Upload
others
View
33
Download
0
Embed Size (px)
Citation preview
-117.5
-116.0-116.5
-117.033.0
34.5
33.5
34.0
5
4
3
2
1
Dep
th (k
m)
Vs
3.2
2.6
2.8
3.0
3.4
3.6
(km/s)
(Zigone et al, 2015)
Vp (km/s)5.4 7.06.25.8 6.65.0
0
1510
5
0
1510
5
0
1510
5
0
1510
5
Dep
th (k
m)
-40 -20 0 20 40 60Fault Normal Distance (km)
Profile 1
Profile 2
Profile 3
Profile 4
(Allam & Ben-Zion, 2012)
Rayleigh Group Velocities (m/s)
Frequency = 3 Hz33.542
33.536
33.538
33.540
-116.595 -116.592 -116.589
250 650450350 550
(Roux et al, 2016)
UTC Time (hours)2 4 6 8 10 12 14 16 18 20 22
Local Time (hours)19 21 23 1 3 5 7 9 11 13 15
50
150
250
50
150
250
Freq
(Hz)
Freq
(Hz)
0
1
2
-2
-1
x 10-2
x 10-3
0
1
2
-2
-1
Correlation Time (s)-0.2 -0.1 0 .01 0.2
10-100 Hz
100-200 Hz
-40000
4000
-50000
5000
Waveforms, Spectrograms & Correlograms
33.75
33.25
33.50
-117.00 -116.75 116.50
(Ben-Zion et al, 2015)
34.5
33.0
33.5
34.0
32.5-117.5 -117.0 -116.5 -116.0 -115.5
-117.5 -117.0 -116.5 -116.0 -115.5
P1P2
P3P4
Hierarchical Seismic Networks around the San Jacinto Fault Zone
2016ANNUAL MEETINGSouthern Ca l i fo rn ia Ear thquake Center
September 10-14, 2016PROCEEDINGS VOLUME XXVI
WELCOME
Core Institutions and Board of Directors (BoD) USC
Tom Jordan, Chair
Harvard
Jim Rice
Texas A&M
Patrick Fulton
UC Santa Barbara
Ralph Archuleta
USGS Menlo Park
Ruth Harris and Steve Hickman
Caltech
Nadia Lapusta, VC
MIT
Tom Herring
UC Los Angeles
Peter Bird
UC Santa Cruz
Emily Brodsky
USGS Pasadena
Rob Graves
CGS
Chris Wills
SDSU
Tom Rockwell
UC Riverside
David Oglesby
UNR
Glenn Biasi
At-Large Member
Roland Bürgmann
Columbia
Bruce Shaw
Stanford
Paul Segall
UC San Diego
Yuri Fialko
USGS Golden
Jill McCarthy
At-Large Member
Michele Cooke
Science Working Groups & Planning Committee (PC) SCEC4
PC Chair
Greg Beroza*
Seismology
Egill Hauksson*
Elizabeth Cochran
Tectonic Geodesy
Dave Sandwell*
Gareth Funning
EQ Geology
Mike Oskin*
Whitney Behr
Computational Sci
Yifeng Cui*
Eric Dunham
PC Vice-Chair
Judi Chester*
USR
John Shaw*
Brad Aagaard
SoSAFE
Kate Scharer*
Ramon Arrowsmith
EFP
Jeanne Hardebeck*
Ilya Zaliapin
EEII
Jack Baker*
Jacobo Bielak
* PC Members FARM
Greg Hirth*
Pablo Ampuero
SDOT
Kaj Johnson*
Liz Hearn
GMP
Kim Olsen*
Domniki Asimaki
CME
Phil Maechling*
CSEP
Max Werner*
Danijel Schorlemmer
WGCEP
Ned Field*
GMSV
Nico Luco
Sanaz Rezaeian
Code Verification
Ruth Harris
SIV
Pablo Ampuero
EQ Simulators
Terry Tullis
Transient Detection
Rowena Lohman
SCEC5
PC Chair
Greg Beroza*
Seismology
Yehuda Ben-Zion*
Jamie Steidl
Tectonic Geodesy
Dave Sandwell*
Gareth Funning
EQ Geology
Mike Oskin*
Whitney Behr
Computational Sci
Eric Dunham*
Ricardo Taborda
PC Vice-Chair
Judi Chester*
FARM
Nadia Lapusta*
Nick Beeler
SDOT
Kaj Johnson*
Bridget Smith-Konter
EFP
Max Werner*
Ned Field
Ground Motions
Domniki Asimaki*
Annemarie Baltay
* PC Members EEII
Jack Baker*
Jon Stewart
SAFS
Kate Scharer*
Michele Cooke
CXM
Liz Hearn*
Brad Aagaard
Special Projects
Christine Goulet*
Phil Maechling*
CEO Planning Committee (CEO PC) * Board liaison
** PC liaison
*** AC liaison
Tim Sellnow***, Chair
U Central Florida
Kate Long***
CalOES
Danielle Sumy
IRIS
Jacobo Bielak**
CMU
Sally McGill
CSUSB
Chris Wills*
CGS
Advisory Council (AC) Gail Atkinson, Chair
Western U
Donna Eberhart-Phillips
UC Davis
M. Meghan Miller
UNAVCO
John Vidale
U Washington
Norm Abrahamson
PG&E
Kate Long
CalOES
Farzad Naeim
John A Martin
Andrew Whittaker
MCEER/Buffalo
Roger Bilham
U Colorado
Warner Marzocchi
INGV Rome
Tim Sellnow
U Central Florida
WELCOME
2016 SCEC Annual Meeting page 3
Table of Contents
SCEC Leadership .......................................................................................................... 1
Table of Contents .......................................................................................................... 3
State of SCEC, 2016 ...................................................................................................... 4
2015 Report of the Advisory Council ......................................................................... 14
Communication, Education, and Outreach Highlights ............................................ 18
Research Accomplishments ...................................................................................... 24
Draft 2017 Science Plan .............................................................................................. 79
Saturday, September 10 ........................................................................................... 112
Plenary Presentations ............................................................................................... 120
Poster Presentations ................................................................................................ 129
Meeting Abstracts ..................................................................................................... 149
Meeting Participants ................................................................................................. 260
Center Management .................................................................................................. 267
SCEC Institutions ...................................................................................................... 267
WELCOME
2016 SCEC Annual Meeting page 4
State of SCEC, 2016
Thomas H. Jordan, SCEC Director
Gregory C. Beroza, SCEC Co-Director
Welcome to the 2016 Annual Meeting!
We welcome you to the 26th Annual Meeting of the Southern California Earthquake Center. Each year of the past
quarter century, the SCEC community has gathered from across the country and around the world to share research
accomplishments and make ambitious science plans. This year, 707 people have pre-registered for the meeting (Figure
1), and 347 poster abstracts have been submitted. The pre-registrants include more than 211 first-time attendees and
almost three hundred undergraduates, graduate students, and postdocs.
During the past year, the SCEC Core Program
has undergone a rigorous five-year review, and
the results have been extremely positive. SCEC
has been authorized by its two principal
sponsoring agencies—the National Science
Foundation and U.S. Geological Survey—for
another five years at a target funding level near
$4.6 million per year. The fifth phase of the
Center (SCEC5) will officially begin on Feb 1,
2017, and continue until January 31, 2022.
The goal of this Annual Meeting is to assess the
progress of our collaborations, refine our draft
science plan, and launch SCEC5 properly by
firing up ambitious research initiatives in new
thematic areas, such as Earthquake Gates and
Beyond Elasticity. We’ll strive, as we do every
year, to learn as much as we can about
earthquakes from the formal presentations and
posters and from the informal discussions with
our scientific colleagues.
To match the bright, clear days forecast for
Palm Springs, the Planning Committee has put together a sparkling program. Saturday and Sunday feature workshops
and discussions on six important topics:
● SCEC SoSAFE Workshop: Recent Successes and Future Challenges
● SCEC Ventura Special Fault Study Area Workshop
● SCEC Workshop on Processes that Control the Strength of Faults and Dynamics of Earthquakes
● SCEC Collaboratory for Interseismic Simulation and Modeling (CISM) Meeting
● SCEC International Workshop on Ground Motion Simulation Validation
● SCEC Workshop on Science Communication: Navigating and Maximizing a Digital, Social World
At 6pm Sunday evening, this year’s Distinguished Speaker, Professor Richard H. Sibson of the University of Otago,
will kick off the main meeting with a plenary lecture on “Earthquakes on Compressional Inversion Structures – Problems
in Mechanics and in Hazard Assessment.” Over the next three days, the agenda will feature keynote speakers
addressing fundamental problems, discussions of major science themes, poster sessions on research results,
earthquake response exercises, technical demonstrations, education and outreach activities, and some lively social
gatherings. The topical titles of the sessions indicate the range of the science we will discuss: Modeling Fault
Systems – Supercycles and Modeling Fault Systems – SCEC Community Models on Monday; Understanding
Earthquake Processes and New Observations and Characterizing Seismic Hazard on Tuesday; Reducing Seismic Risk
on Wednesday. In all of these activities, we value your participation as an active member of the SCEC community!
WELCOME
2016 SCEC Annual Meeting page 5
Assessing SCEC Accomplishments
The SCEC4 science plan was posed in terms of the “six fundamental problems of earthquake science” (Table 1). Over
the past five years, we have collectively approached these interrelated, system-level problems with the interdisciplinary,
multi-institutional research at which the SCEC community excels. This year, SCEC Co-Director Greg Beroza and the
Planning Committee (PC) will be assembling the final report on the SCEC4 research accomplishments, which will be
submitted in spring of 2017 as part of a final report on the Center’s fourth phase to the NSF and USGS. The PC’s draft
report is included in these Proceedings. Greg will summarize the research results, with an emphasis on our more recent
accomplishments, in his plenary address on Monday morning. This meeting volume also contains a report by Mark
Benthien, the SCEC Associate Director for Communication, Education, and Outreach (CEO), on the remarkable
accomplishments of the CEO program.
Table 1. Fundamental Problems of Earthquake Science (SCEC4)
1. Stress transfer from plate motion to crustal faults: long-term fault slip rates 2. Stress-mediated fault interactions and earthquake clustering: evaluation of mechanisms 3. Evolution of fault resistance during seismic slip: scale-appropriate laws for rupture modeling 4. Structure and evolution of fault zones and systems: relation to earthquake physics 5. Causes and effects of transient deformations: slow slip events and tectonic tremor 6. Seismic wave generation and scattering: prediction of strong ground motions
The five poster sessions scheduled between Sunday evening and Tuesday evening will display the entire spectrum of
SCEC accomplishments. Posters will stay up for the entire meeting to allow more face-to-face interactions on the nitty-
gritty aspects of SCEC scientific research.
The SCEC5 Community Science Vision
The proposed SCEC5 Science Plan was developed by the non-USGS members of the SCEC Planning Committee and
Board of Directors with extensive input from issue-oriented “tiger teams” and the community at large. The tiger teams
organized research ideas and plans from the SCEC community into white papers on a number of the most compelling
topics. An ad hoc committee, appointed by the Board and chaired by P. Segall, abstracted from this and other input a
strategic framework for prioritizing SCEC5 research objectives, which has been cast in terms of five basic questions of
earthquake science (Table 2).
Table 2. Basic Questions of Earthquake Science (SCEC5)
1. How are faults loaded across temporal and spatial scales? 2. What is the role of off-fault inelastic deformation on strain accumulation, dynamic rupture, and
radiated seismic energy? 3. How do the evolving structure, composition and physical properties of fault zones and
surrounding rock affect shear resistance to seismic and aseismic slip? 4. How do strong ground motions depend on the complexities and nonlinearities of dynamic
earthquake systems? 5. In what ways can system-specific studies enhance our general understanding of earthquake
predictability?
Science Plan. Research priorities have been developed to address these five basic questions. Tied to the priorities
are fourteen science topics distributed across four main thematic areas.
Modeling the fault system: We seek to know more about the geometry of the San Andreas system as a complex network
of faults, how stresses acting within this network drive the deformation that leads to fault rupture, and how this system
evolves on time scales ranging from milliseconds to millions of years.
- Stress and Deformation Over Time. We will build alternative models of the stress state and its evolution during seismic
cycles, compare the models with observations, and assess their epistemic uncertainties, particularly in the
representation of fault-system rheology and tectonic forcing.
- Special Fault Study Areas: Focus on Earthquake Gates. “Earthquake gates” are regions of fault complexity
conjectured to inhibit propagating ruptures, owing to dynamic conditions set up by proximal fault geometry, distributed
deformation, and earthquake history. We will test the hypothesis that earthquake gates control the probability of large,
multi-segment and multi-fault ruptures.
WELCOME
2016 SCEC Annual Meeting page 6
- Community Models. We will enhance the accessibility of the SCEC Community Models, including the model
uncertainties. Community thermal and rheological models will be developed.
- Data Intensive Computing. We will develop methods for signal detection and identification that scale efficiently with
data size, which we will apply to key problems of Earth structure and nanoseismic activity.
Understanding earthquake processes: Many important achievements in understanding fault-system stresses, fault
ruptures, and seismic waves have been based on the elastic approximation, but new problems motivate us to move
beyond elasticity in the investigation of earthquake processes.
- Beyond Elasticity. We will test hypotheses about inelastic fault-system behavior against geologic, geodetic, and
seismic data, refine them through dynamic modeling across a wide range of spatiotemporal scales, and assess their
implications for seismic hazard analysis.
- Modeling Earthquake Source Processes. We will combine co-seismic dynamic rupture models with inter-seismic
earthquake simulators to achieve a multi-cycle simulation capability that can account for slip history, inertial effects,
fault-zone complexity, realistic fault geometry, and realistic loading.
- Ground Motion Simulation. We will validate ground-motion simulations, improve their accuracy by incorporating
nonlinear rock and soil response, and integrate dynamic rupture models with wave-scattering and attenuation models.
We seek simulation capabilities that span the main engineering band, 0.1-10 Hz.
- Induced Seismicity. We will develop detection methods for low magnitude earthquakes, participate in the building of
hydrological models for special study sites, and develop and test mechanistic and empirical models of anthropogenic
earthquakes within Southern California.
Characterizing seismic hazards: We seek to characterize seismic hazards across a wide spectrum of anticipation and
response times, with emphasis on the proper assessment of model uncertainties and the use of physics-based methods
to lower those uncertainties.
- Probabilistic Seismic Hazard Analysis. We will attempt to reduce the uncertainty in PSHA through physics-based
earthquake rupture forecasts and ground-motion models. A special focus will be on reducing the epistemic uncertainty
in shaking intensities due to 3D along-path structure.
- Operational Earthquake Forecasting. We will conduct fundamental research on earthquake predictability, develop
physics-based forecasting models in the new Collaboratory for Interseismic Simulation and Modeling, and coordinate
the Working Group on California Earthquake Probabilities.
- Earthquake Early Warning. We will develop methods to infer rupture parameters from time-limited data, ground-motion
predictions that account for directivity, basin, and other 3D effects, and better long-term and short-term earthquake
rupture forecasts for conditioning of early-warning algorithms.
- Post-Earthquake Rapid Response. We will improve the rapid scientific response to strong earthquakes in Southern
California through the development of new methods for mobilizing and coordinating the core geoscience disciplines in
the gathering and preservation of perishable earthquake data.
Reducing seismic risk: Through partnerships coordinated by SCEC’s Earthquake Engineering Implementation
Interface, we will conduct research useful to motivating societal actions to reduce earthquake risk. Two topics
investigated by these engineering partnerships will be:
- Risk to Distributed Infrastructure. We will work with engineers and stakeholders to apply measures of distributed
infrastructure impacts in assessing correlated damage from physics-based ground-motion simulations. An initial project
will develop earthquake scenarios for the Los Angeles water supply.
- Earthquake Physics of the Geotechnical Layer. In collaboration with geotechnical engineers, we will advance the
understanding of site effects and soil-structure interactions by incorporating nonlinear rheological models of near-
surface rock and soil layers into full-physics earthquake simulations.
The Planning Committee has synthesized a draft of the SCEC5 science plan, which is included in the Meeting
Proceedings. At this annual meeting, we are soliciting input from the entire SCEC community on the details of this plan.
A revised version derived from this input will be posted in early October, which will be the basis for a request for
proposals, due in early November.
WELCOME
2016 SCEC Annual Meeting page 7
Communication, Education and Outreach Plan. The SCEC/CEO program will manage and expand a suite of
successful activities within four CEO focus areas. Knowledge Implementation will connect SCEC scientists and
research results with practicing engineers, government officials, business risk managers, and other professionals in
order to improve application of earthquake science. The Public Education and Preparedness focus area will educate
people of all ages about earthquakes, tsunamis, and other hazards, and motivate them to become prepared. The K-14
Earthquake Education Initiative will improve earth science education in multiple learning environments, overall science
literacy, and earthquake safety in schools and museums. The Experiential Learning and Career Advancement program
will provide research opportunities, networking, and other resources to encourage students and sustain careers in
STEM fields. Four long-term intended outcomes of the CEO program are improved application of earthquake science
in policy and practice; reduced loss of life, property, and recovery time; increased science literacy; and increased
diversity, retention, and career success in the scientific workforce. SCEC’s vigorous promotion of workforce diversity
will be augmented by a new Transitions Program that will provide students and early-career scientists with resources
and mentoring at major steps in their careers.
Requests from the Sponsoring Agencies. The SCEC5 proposal process is not over yet. In order to advance NSF’s
formal recommendation, we must submit a revised work plan and budget to our NSF program officer, Greg Anderson,
by 15 October 2016. Anderson’s letter specifically requests:
1. A clear and specific statement of the impacts of the NSF budget reduction from $4.1M (requested) to $3.0M
(authorized), and the differences between the proposal and the revised request.
2. A work plan, including milestones for each of the four key themes of the SCEC5 Science Plan, which should be
relatively detailed for the first year and less so in subsequent years. The milestones will form one basis of the annual
project report and may be revised as part of the annual collaboration process in future years.
3. A clear description of a revised CEO evaluation framework based on the CEO logic model, which will form another
basis of the annual project report.
4. A plan to keep NSF apprised of progress in carrying out the center leadership transition.
The review by the USGS Earthquake Hazard Program (EHP) focused on the intersection of SCEC plans with projected
USGS activities. The letter from our contracting officer, Margaret Eastman, requested SCEC’s consideration of several
EHP priorities, which include: (a) priority on community-model development that will serve broader research needs and
stimulate model developments elsewhere in the country; (b) coordination requirements on earthquake early warning
and induced-seismicity research; (c) clarification on how SCEC’s earthquake response planning and activities will be
coordinated with the USGS and with the California Clearinghouse; (d) partnership with the USGS in earthquake
engineering implementation and in interactions with code committees and design teams. The EHP explicitly
encouraged potential efforts to help translate SCEC results into earthquake system science, hazards assessment, and
engineering practice outside of Southern California.
We must submit our revised plans to the agencies within the next month. At this meeting, the SCEC Board of Directors
and the Planning Committee will be considering how we should best respond to these requests, and we would value
your input.
Organization and Leadership
SCEC has developed an effective management structure for coordinating earthquake research and educational
activities. The Center’s ability to facilitate collaborative, investigator-driven research has been repeatedly proven by its
diverse accomplishments. Participation in SCEC is rising despite flat funding (Figure 1), and its national and
international partnerships are flourishing. In its annual reports, the SCEC External Advisory Council has repeatedly
documented the enthusiasm among SCEC participants and endorsed their high levels of satisfaction with the Center’s
leadership and administration.
In preparing the core-program proposal, the SCEC Board of Directors voted unanimously to operate SCEC5 under a
similar set of by-laws as SCEC4. The University of Southern California (USC) will continue as the managing institution,
and Tom Jordan, the proposal PI, will continue as the Center Director. The by-laws now designate the responsibilities
of a Center Co-Director, Greg Beroza of Stanford University, who is Co-PI on the SCEC5 proposal. Establishment of a
co-directorship and several other augmentations to the SCEC leadership structure have been designed to facilitate the
SCEC leadership transition, which we anticipate will occur early in SCEC5.
WELCOME
2016 SCEC Annual Meeting page 8
Core and Participating Institutions. SCEC will continue as an institution-based center, governed by a Board of
Directors, who represent its members. The Center currently involves more than 1000 scientists and other experts in
active SCEC projects, making it one of the largest formal collaborations in geoscience. It will continue to operate as an
open consortium, available to all qualified individuals and institutions seeking to collaborate on earthquake science in
Southern California, and its membership will continue to evolve. The institutional membership currently stands at 75,
comprising 18 core institutions and 57 participating institutions, which are listed on the inside back cover of the meeting
program. As you can see from the list, SCEC institutions are not limited to universities, nor to U.S. organizations. The
three USGS offices in Menlo Park, Pasadena, and Golden and the California Geological Survey are core institutions,
and AECOM Corporation is a participating institution. Twelve foreign institutions are currently recognized as partners
with SCEC through a set of international cooperative agreements.
Call for Participating Institutions. All SCEC4 participating institutions, as well as institutions that would like to join the
Center, are requested to apply for institutional membership in SCEC5 before December 31, 2016. The process is an
easy one; all we need is a letter from a cognizant official (e.g., your department chair or dean) that requests
participating-institution status and appoints an institutional representative who will act as the point-of-contact with the
Center.
Board of Directors. The complete SCEC4 Board of Directors, which now includes Texas A&M as a core institution, is
listed on the inside front cover of the meeting program. Each core institution will appoint one member to the SCEC5
Board of Directors, which will be chaired by the Center Director. The Board will elect two nominees from the participating
institutions to serve two-year terms as members-at-large.
Nominations are now open for the at-large members of the SCEC5 Board.
The Board will be the primary decision-making body of SCEC; it will meet three times per year (typically in February,
June, and September) to approve the Annual Collaboration Plan and budget and deal with major business items. Based
on current projections, the SCEC5 Board will comprise 17 voting members. The USGS members will serve in non-
voting liaison capacity. Ex officio members will include the Co-Director; the Associate Director for Administration
(serving as Executive Secretary); the Associate Director for CEO; the IT Architect; and the Executive Science Director
for Special Projects.
An Executive Committee will handle daily decision-making responsibilities, mainly through email. It will comprise five
voting members, the Center Director, who will act as Chair, the Co-Director, the Board Vice-Chair, and two Board
members elected for 3-year terms, plus three non-voting members: the Executive Director for Special Projects, the AD
for CEO, and the AD for Administration.
External Advisory Council. The external Advisory Council (AC) will continue to serve as an experienced advisory
body to the Center, charged with developing an overview of SCEC operations and advising the Director and the Board.
Since the inception of SCEC in 1991, the AC has played a major role in maintaining the vitality of the SCEC and helping
its leadership chart new directions. The Center has always provided its sponsoring agencies and participants, with
verbatim copies of the yearly AC reports. The full 2015 AC report is included in this volume. The current AC membership
can be found in the meeting program.
We are very happy to announce that Professor John Vidale of the University of Washington has agreed to chair the
SCEC5 AC, filling the big shoes of Professor Gail Atkinson of Western University, who will step down from this role at
the end of SCEC4. We will have a chance at the annual banquet on Monday evening to thank Gail for her excellent
leadership of the AC and to remind her of continuing role as a denizen of Hotel California.
Planning Committee. The chair of the Planning Committee (PC) is the SCEC Co-Director, Greg Beroza of Stanford,
and its Vice-Chair is Judi Chester of Texas A&M. The PC comprises the leaders of the SCEC science working groups—
disciplinary committees, focus groups, and special project groups—who, together with the working group co-leaders,
guide SCEC’s research program. The PC is responsible for formulating the Center’s science plan, conducting proposal
reviews, and recommending projects to the Board for SCEC support. Its members will play key roles in implementing
the SCEC5 science plan. To prepare for SCEC5, we have restructured the PC working groups and evolved its
membership. The “Super-PC”, comprising the SCEC4 PC members as well as the new SCEC5 PC members, are here
WELCOME
2016 SCEC Annual Meeting page 9
at the meeting; their names are listed on inside
cover of the meeting program. We urge you to
use the opportunity of the Annual Meeting to
communicate your thoughts about future
research plans to them.
Working Groups. The SCEC organization
comprises a number of disciplinary
committees, focus groups, special project
teams, and technical activity groups (TAGs).
These working groups have been our engines
of success, and many of the discussions at
this meeting will feed into their plans.
The Center supports disciplinary science
through standing committees in Seismology,
Tectonic Geodesy, Earthquake Geology,
and Computational Science. These groups
(green boxes of Figure 2) are responsible for
disciplinary activities relevant to the SCEC
Science Plan, and they make
recommendations to the Planning
Committee from the perspective of disciplinary research and infrastructure. The groups are unchanged in the transition
to SCEC5, though there are changes in leadership. Seismology: Yehuda Ben-Zion and Jamie Steidl; Tectonic Geodesy:
Dave Sandwell and Gareth Funning; Earthquake Geology: Mike Oskin and Whitney Behr; Computational Science:Eric
Dunham and Ricardo Taborda.
SCEC coordinates earthquake system science through interdisciplinary focus groups. The SCEC4 focus groups
included Unified Structural Representation (USR), Fault and Rupture Mechanics (FARM), Earthquake Forecasting and
Predictability (EFP), Southern San Andreas Fault Evaluation (SoSAFE), Stress and Deformation Through Time
(SDOT), Ground Motion Prediction (GMP), and the Earthquake Engineering Implementation Interface (EEII). Most
interdisciplinary groups will continue in SCEC5, with some changes in leadership. FARM: Nadia Lapusta and Nick
Beeler; EFP: Max Werner and Ned Field; SDOT: Kaj Johnson and Bridget Smith-Konter; EEII: Jack Baker and Jon
Stewart. GMP (Kim Olsen, Domniki Asimaki) becomes “Ground Motions” (GM) in SCEC5 (Domniki Asimaki, Annemarie
Baltay-Sundstrom). The SoSAFE working group will evolve into the San Andreas Fault System (SAFS) working group,
led by Kate Scharer and Michele Cooke, with a greater emphasis on modeling the fault system. USR will be broadened
into a CXM focus group, led by Liz Hearn and Brad Aagaard, and encompass activity related to the constructions,
improvement, and maintenance of all community model types, as described in the SCEC5 proposal. The importance
and scale of effort involved with community models led us to request funding for a Community Models Manager in the
SCEC5 proposal. The requested budget increase was not granted, however, which will present challenges in this
important area of SCEC5 research.
Technical Activity Groups are self-organized to develop and test critical methodologies for solving specific problems.
TAGs have formed to verify the complex computer calculations needed for wave propagation and dynamic rupture
problems, to assess the accuracy and resolving power of source inversions, and to develop geodetic transient detectors
and earthquake simulators. TAGs share a modus operandi: the posing of well-defined “standard problems”, solution of
these problems by different researchers using alternative algorithms or codes, a common cyberspace for comparing
solutions, and meetings to discuss discrepancies and potential improvements. Existing TAGs in SCEC4 will sunset.
TAGs in these and other areas can be reinitiated at the beginning of SCEC5 through successful proposals to the PC.
Special Projects
The SCEC special projects are research partnerships in targeted earthquake research that heavily leverage the core
program. Synergy between the special projects and the core program is ensured by a central SCEC policy, instituted
by the Board of Directors in 2005: the science objectives of all SCEC special projects must be aligned with those of the
SCEC core program and explicitly included as objectives in the SCEC Annual Science Plan. Under this policy, any
WELCOME
2016 SCEC Annual Meeting page 10
SCEC participant can propose core-program research pertinent to a special project, enabling them to participate in that
project.
Community Modeling Environment (CME). The CME is SCEC’s high-performance computing collaboratory for large-
scale earthquake simulations. Major grants to support CME software engineering have come from the NSF/CISE
Directorate and the NSF/EAR Geoinformatics program, as well as from the utility industry. SCEC competes for
supercomputer allocations through the NSF XSEDE and PRAC programs and the DOE INCITE program. In 2015,
SCEC was awarded allocations totaling 362 million service units, primarily on the NCSA’s Blue Waters, ANL’s Mira,
and ORNL’s Titan supercomputers. These resources have enabled SCEC to sustain its HPC usage at a high level of
productivity. CME resources support five major SCEC computational platforms:
High-F Platform: The High-F platform comprises the AWP-ODC, Hercules, and other codes that SCEC researchers
are using to push earthquake simulations to higher frequencies (> 1 Hz). Software under development will be
capable of modeling the effects of fault roughness, near-fault plasticity, frequency-dependent attenuation,
topography, small-scale near-surface heterogeneities, and near-surface nonlinearities. The High-F Platform will
support dynamic-rupture and ground-motion studies as part of the SCEC5 plans to move simulations Beyond
Elasticity.
CyberShake Platform: The CyberShake Platform uses seismic reciprocity to generate large ensembles of
simulations (> 108) that Monte-Carlo sample earthquake rupture forecasts and multiple crustal-structure models.
Implementation of physics-based probabilistic seismic hazard modeling requires this capability. The platform is
being developed using the Los Angeles region as a test bed, and it has already produced PSHA models as
candidates for the USGS Urban Seismic Hazard Mapping Project. In SCEC5, CyberShake PSHA models will be
developed for other regions; e.g., in Central California as part of the Central California Seismic Project (see below).
Because reciprocity derives from linear elasticity, a SCEC5 challenge will be the reengineering of CyberShake to
enable the efficient, large-ensemble simulation of nonlinear wave phenomena that lie Beyond Elasticity.
Broadband Platform: The open-source Broadband Platform (BBP) provides a verified, validated, and user-friendly
computational environment for generating broadband (0-100Hz) ground motions. In its validation mode, the BBP
computes goodness-of-fit measures that quantify how well the synthetics match the observations. In its scenario
mode, it calculates suites of synthetic seismograms from user-specified rupture sets, structural models, and station
sets. In SCEC5, the BBP will be extended from 1D to 3D structural models, and it will support the development and
validation of physics-based ground-motion models in projects and partnerships managed under the EEII.
F3DT Platform: This platform integrates the software needed for full-3D waveform tomography using the adjoint-
wavefield and scattering-integral formulations of the structural inverse problem. F3DT can invert both earthquake
waveforms and ambient-field correlograms for high-resolution crustal models, and it can refine the centroid moment
tensors of earthquakes by matching observed waveforms with 3D synthetics. These capabilities have been used to
produce CVM-S4.26 and will be employed in the SCEC5 CVM studies.
Unified Community Velocity Model Platform. The UCVM platform provides an easy-to-use software framework for
comparing and synthesizing 3D Earth models and delivering model products to users. This community software is
an important component of the CME cyberinfrastructure: a standardized, high-speed query interface enables users
to build very large simulation meshes very quickly, and its file utilities can export meshes in both eTree and NetCDF
formats.
Uniform California Earthquake Rupture Forecast. UCERF is a joint project of SCEC, USGS, and CGS to build a
California-wide, time-dependent, fault-based earthquake rupture forecast, managed through the Working Group on
California Earthquake Probabilities. The latest (third) version comprises a time-independent model used in the 2014
release of the National Seismic Hazard Mapping Project (UCERF3-TI), a time-dependent model based on long-term
Reid renewal statistics (UCERF3-TD), and a time-dependent model based on short-term ETAS statistics (UCERF3-
ETAS). The latter is being developed as a candidate model for use in operational earthquake forecasting. CSEP testing
of UCERF3-ETAS will commence in 2016. A major SCEC5 initiative is to incorporate more physics into UCERF models
through the use of physics-based earthquake simulators.
Collaboratory for the Study of Earthquake Predictability. CSEP provides an international cyberinfrastructure that
sustains the prospective, blind testing of short- and medium-term earthquake forecasts on regional and global scales.
CSEP testing centers (except Japan and China) run a common software stack that is developed and released quarterly
WELCOME
2016 SCEC Annual Meeting page 11
by SCEC software engineer Maria Liukis. CSEP operations at SCEC include the testing of California and global
forecasting models in addition to the development and maintenance of the collaboratory software. SCEC is responsible
for the registration of new models and the coding of new testing procedures; many of these innovations, such as the
testing of geodetic anomaly detectors, have come from the SCEC core program. Over 400 earthquake-forecasting
models and their variations are currently under prospective CSEP testing. The registration and testing of external
forecasting models is underway and may eventually include USGS operational models. CSEP and its new sister
collaboratory, CISM (see below), will be critical SCEC5 infrastructures for the development and evaluation of
comprehensive earthquake forecasting models. The proposed earthquake-simulator effort is supported by the recent
results of the CSEP Canterbury Retrospective Experiment, which has demonstrated (in one particular sequence) that
models incorporating the physics of rate-state nucleation and Coulomb stress transfer can outperform purely statistical
models such as ETAS.
CSEP was initiated in 2006 with support from the W. M. Keck Foundation and has been subsequently funded under
grants and contracts from the USGS and Department of Homeland Security. Owing to CSEP’s importance to SCEC
core research, the SCEC5 proposal requested USGS funding of $200K/year to support collaboratory operations in
California and USGS-relevant software development, and this budgetary increase has been approved. Allocation of
CSEP resources to specific projects will be guided as part of the core budgeting process by the Joint SCEC/USGS
Planning Committee.
Collaboratory for Interseismic Simulation and Modeling. In July, 2015, SCEC received a three-year, $2M grant from the
W. M. Keck Foundation to construct a Collaboratory for Interseismic Simulation and Modeling. CISM will provide a
unique environment for developing large-scale numerical models that can simulate sequences of fault ruptures and the
seismic shaking they produce. The goal of CISM is to equip earthquake scientists with HPC-enabled infrastructure for
creating a new generation of comprehensive, physics-based earthquake forecasts using California as the primary test
bed. CISM will provide a computational framework for combining earthquake simulations that account for the physics
of earthquake nucleation and stress transfer with ground-motion simulations. It will be engineered as a workflow-
oriented cyberinfrastructure with common tools for integrating various types of scientific software modules provided by
different research teams into well-structured forecasting models that can be calibrated against existing data and tested
against observations within CSEP. As part of this project, W. M. Keck Foundation Fellowships in Earthquake
Forecasting Research will support participation in CISM by graduate students, post-docs, and early-career researchers.
The first of the CISM postdoctoral fellowships was awarded to Dr. Jacqui Gilchrist in March 2016.
Central California Seismic Project. The CCSP was initiated in 2015 as a partnership between SCEC and Pacific Gas
& Electric Co. to use the central coast region of California as a testbed for developing and validating new physics-based
ground-motion forecasting models. The main goal of the CCSP is to assess the effectiveness of seismic wavefield
modeling in reducing the epistemic uncertainties in path effects that control hazard estimates at low exceedance
probabilities. The specific objectives of this long-term (~8-yr) effort include (i) assimilation of existing data into improved
3D models of Central California crustal structure; (ii) collection of new data on local earthquake activity and regional
path effects, (iii) validation of improvements to synthetic seismograms derived from 3D models, and (iv) demonstration
that physics-based modeling can reduce path-effect uncertainties. Work on objective (i) has begun, and a start has
been made on the instrument deployments required to achieve (ii). The CCSP will provide SCEC researchers with new
data in both location and type, and its objectives are well aligned with the SCEC5 Basic Questions of Earthquake
Science, especially Q4 and Q5 (Table 1).
Mining Seismic Wavefields Project. In May, 2016, a SCEC working group led by PI Greg Beroza (Stanford) and co-PIs:
Zhigang Peng (Georgia Tech), Egill Hauksson (Caltech), Yehuda Ben-Zion (USC), Phil Maechling (USC), and Tom
Jordan (USC), received a two-year grant from the NSF geoinformatics program to develop and deploy
cyberinfrastructure for mining seismic wavefields through data intensive computing techniques to extend similarity
search for earthquake detection to massive data sets. Similarity search has been used to understand the mechanics
of tectonic tremor, transform our understanding of the depth-dependence of faulting, illuminate diffusion within
aftershock seismicity, and reveal new insights into induced earthquakes. These results were achieved with modest
data volumes – from ~ 10 seismic stations spanning ~ 10 km – yet they increased the number of detected earthquakes
by a factor of 10 to 100. This geoinformatics project will develop the cyberinfrastructure required to enable high-
sensitivity studies of earthquake processes through the discovery of previously undetected seismic events within both
long duration (large-T) and instrumentally dense (large-N) data sets.
WELCOME
2016 SCEC Annual Meeting page 12
Center Budget and Project Funding
The SCEC base program has been essentially flat funded by NSF/EAR and USGS/EHP since the beginning of SCEC2
(Figure 1). In 2013 NSF/EAR cut our base budget by 10%, from the $3.0M received in 2012 to $2.7M. The USGS cuts
were proportionately smaller (3%), from $1.34M to $1.30M. A similar cut of 10% was expected in 2014, a result of the
federal sequester law and congressional funding levels, but NSF announced late in the year that SCEC would get its
full $3.0M in 2014. The USGS did continue its small 3% cut. In 2015, we received $2.9M from NSF and $1.3M from the
USGS, both cuts of ~3%. In 2016, NSF/EAR cut base funding by $66K (from the $3.0M per year authorized SCEC4
level) to $2.924M. The USGS fully funded SCEC at the authorized level of $1.34M in 2016. We can report NSF/EAR is
making a late supplement of $40K to help support the increased participation by graduate (122 to 190) and
undergraduate (12 to 40) students at this year’s annual meeting. Supplementing the $4.304M in base funding is $500K
from Pacific Gas & Electric, the Keck Foundation, and the geodesy royalty fund. In total, SCEC core funding for 2016
is $4,944K, up slightly from $4,890K in 2015.
Our 2016 funding was not finalized until May, but the SCEC administration team was able to work around the delay to
get nearly all awards out in a timely manner for the final year of SCEC4 funding. The team continues to monitor
subcontracts closely to prevent any major closeout problems at the end of SCEC4 on January 31, 2017.
The base budget approved by the Board of Directors for this year allocated $3.42M (versus $3.384M in 2015) for
science activities managed by the SCEC Planning Committee; $470K for communication, education, and outreach
activities, managed by the CEO Associate Director, Mark Benthien; $190K for information technology, managed by
Associate Director for Information Technology, Phil Maechling; $400K for administration and $380K for meetings,
managed by the Associate Director for Administration, John McRaney; and $64K for the Director's reserve account.
The latter funding was reduced from $130K due to the original cut by NSF.
Structuring of the SCEC program for 2016 began with the working-group discussions at our last Annual Meeting in
September, 2015. An RFP was issued in October, 2015, and 161 proposals (~225 individual funding requests counting
collaborative proposals) requesting a total of $4.86M were submitted in November, 2015. Both the number of proposals
and the total funds requested were slightly lower than those submitted in 2014 for the 2015 SCEC science program.
The small drop was expected since the RFP did not allow new project starts in the last year of SCEC4. All proposals
were independently reviewed by the Director and by either the PC Chair or Vice-Chair. Each proposal was also
independently reviewed by the leaders and/or co-leaders of three relevant focus groups or disciplinary committees.
(Reviewers were required to recuse themselves when they had a conflict of interest.) The PC met in January 2016, and
spent two days discussing every proposal. The objective was to formulate a coherent, budget-balanced science
program consistent with SCEC's basic mission, short-term objectives, long-term goals, and institutional composition.
Proposals were evaluated according to the following criteria:
● Scientific merit of the proposed research
● Competence and performance of the investigators, especially in regard to past SCEC-sponsored research
● Priority of the proposed project for short-term SCEC objectives as stated in the RFP
● Promise of the proposed project for contributing to long-term SCEC goals as reflected in the SCEC science
plan
● Commitment of the investigator and institution to the SCEC mission
● Value of the proposed research relative to its cost
● Ability to leverage the cost of the proposed research through other funding sources
● Involvement of students and junior investigators
● Involvement of women and underrepresented groups
● Innovative or "risky" ideas that have a reasonable chance of leading to new insights or advances in earthquake
physics and/or seismic hazard analysis.
● The need to achieve a balanced budget while maintaining a reasonable level of scientific continuity
The recommendations of the PC were reviewed by the SCEC Board of Directors. The Board voted unanimously to
accept the PC's recommendations. After minor adjustments and a review of the proposed program by the NSF and
USGS, the Director approved the final program in late February 2016. The science plan was then sent to our NSF and
USGS program officers for agency review. Once that approval was received, notifications to investigators were sent,
starting in March 2016.
WELCOME
2016 SCEC Annual Meeting page 13
Communication, Education, and Outreach
The success of SCEC’s CEO program matches that of its science program. CEO offers a wide range of student
research experiences, web-based education tools, classroom curricula, museum displays, public information
brochures, online newsletters, workshops, and technical publications. Highlights of CEO activities for the past year are
reported in these Proceedings by the Associate Director for CEO, Mark Benthien, who will present an oral summary on
Monday morning.
In 2015, we established a CEO Planning Committee with members selected to represent the four CEO focus areas.
The CEO-PC was chartered to provide guidance and support for the portfolio of SCEC/CEO activities and partnerships,
review reports and evaluations, and identify synergies with other parts of SCEC and external organizations. The CEO-
PC was convened in Spring 2015 and in June 2016 with members drawn from the AC and SCEC stakeholders. The
Chair of the CEO-PC is Tim Sellnow (U. Central Florida), and the Vice-Chair is Kate Long (CalOES). Both represent
the Public Education and Preparedness CEO focus area and are also on the AC. Danielle Sumy (IRIS) represents K-
14 Earthquake Education Initiative. Sally McGill (CSU San Bernardino) represents the Experiential Learning and Career
Advancement focus area. Jacobo Bielak (Carnegie Mellon University) and Chris Wills (California Geological Survey)
both represent the Implementation Interface CEO focus area. Chris is also the representative of the SCEC Board on
the CEO-PC.
A Special Word of Thanks
SCEC has been successful because of the collaborative efforts of many people over many years. We want to express
our deep appreciation to all of you for your attendance at the Annual Meeting and your sustained commitment to the
collaboration. The SCEC5 proposal was a vision statement by the entire community, and we thank the many SCEC
participants, especially to the PC and Board members, who spent substantial time contributing to this successful
document.
Special recognition is in order for SCEC staff, which comprises individuals of remarkable skills and dedication. We all
benefit immensely from the financial wizardry and personal empathy of John “The Chaplain” McRaney, the
organizational skills of Mark “Mr. ShakeOut” Benthien, and the innovative expertise of Phil “Big-Iron” Maechling. We
now also benefit from the outstanding scientific and organizational talents of Christine Goulet, who has assumed with
full vigor the many duties of the SCEC Executive Director for Special Projects.
And we all owe a very special thank you to Tran Huynh and Deborah Gormley, the SCEC Meetings Team, and their
diligent associates, Karen Young, Edric Pauk, John Marquis, David Gill, and Jason Ballmann, for their exceptional
efforts in organizing this meeting and arranging its many moving parts. Please do not hesitate to contact us, Tran, or
other members of the SCEC team if you have questions or comments about our meeting activities and future plans.
Now please enjoy the sessions, the meals, and the pool in the spectacular and sparkling setting of Palm Springs!
WELCOME
2016 SCEC Annual Meeting page 14
2015 Report of the Advisory Council
Gail Atkinson, SCEC Advisory Council Chair
Introduction
The SCEC Advisory Committee (AC) met at the Annual SCEC meeting in Palm Springs from Sept. 13 to 16, 2015 to
review SCEC activities and offer advice to the SCEC leadership. The SCEC AC comprises the following members
(names indicated with * are members who were present at the meeting):
Gail Atkinson*, Chair (Western University) [email protected]
Norm Abrahamson*(Pacific Gas and Electric)
Roger Bilham* (University of Colorado)
Donna Eberhart-Phillips* (UC Davis)
Kate Long* (California Office of Emergency Services)
Warner Marzocchi* (INGV, Rome)
M. Meghan Miller* (UNAVCO)
Farzad Naeim (Farzad Naeim Inc.)
Tim Sellnow* (University of Kentucky)
John Vidale* (University of Washington)
Andrew Whittaker (University of Buffalo)
The AC met initially on Sept. 13 and was briefed by the SCEC leadership. Director Jordan provided the AC with a
summary of the state of SCEC and posed a list of issues on which AC feedback was sought. Following the leadership
briefing, the AC discussed the agenda for the next few days and shared initial thoughts. The key focus activities for this
meeting were defined at that time as: (i) a review of SCEC4 accomplishments, and any suggestions for areas to focus
efforts in the final year of SCEC4; and (ii) an overview-level review of the SCEC5 proposal draft (to be submitted to
funding agencies by SCEC no later than Oct. 1). The purpose of the AC preview of the SCEC5 proposal was to provide
confidential feedback to SCEC leadership for their consideration in fine-tuning the final proposal. That information was
conveyed separately to the SCEC leadership and is not a part of this report.
Over the following three days, the AC attended scientific sessions and solicited impressions and feedback from
attendees. A session with the SCEC/CEO team under Associate Director Benthien was held Monday, and two members
of the AC also participated in the CEO Planning Committee meeting on Tues. evening. The AC also reviewed a
comprehensive workbook prepared for us by the SCEC leadership, as well as reviewing a draft of the SCEC5 proposal.
The AC reconvened Tues. mid-day and Tues. evening to compile their report and recommendations, which was
presented to the SCEC community on Wed. morning.
Our overall impression is that over its 25 year history, SCEC has become the world’s most effective, sustained and
cohesive collaboration of earthquake scientists, dedicated to understanding the physics behind earthquake hazards at
all scales, and addressing their impacts on society. SCEC has international stature and recognition as a model of the
benefits of collaboration, wherein the whole is greater than the sum of its parts. This is all the more remarkable because
the SCEC parts represent a stunning breadth of expertise. SCEC displays consistently cutting-edge science, making
major inroads in understanding earthquake faulting processes and their implications for ground motions. SCEC’s
earthquake engineering interactions represent a major SCEC4 accomplishment that provides a compelling rationale
for support of SCEC5.
We discussed the specific issues and questions posed to us by SCEC Director Jordan, and offer the following
comments and observations.
Structure of the Advisory Committee
Director Jordan requested our input on whether the structure of the AC is effective, and solicited ideas on recruiting
new members. We believe that the AC structure works well and we do not suggest any changes. Recruiting new and
continued engineering participation would be useful. One possibility would be to tap into the globally-oriented engineers
in groups such as those coordinated by Brian Tucker or Elizabeth Hausler (those individuals might be asked for
WELCOME
2016 SCEC Annual Meeting page 15
suggestions?). It may also be useful to solicit participation of an LA-based engineer. On the simulations side, an AC
member with knowledge in earthquake physics would be helpful.
SCEC Management Structure
The AC is very satisfied with the new management structure. We think that a new search for the next Director should
begin in 2016-2017, following initiation of SCEC5. Replacing Tom Jordan will be a tall order, even with the helpful
changes to the management structure that have been made. It will be equally important and challenging to replace
John McRaney. The AC would like to stress to USC the importance to the future success of SCEC of a timely and
fruitful search for both of these positions. This may require some flexibility on the part of USC in regards to balancing
their ideal qualifications and conditions against the practical realities of directing a major organization like SCEC.
Annual Meeting, and Engagement of New Scientists
SCEC could consider surveying early-career level scientists at the next meeting, and asking for their suggestions on
how to best enhance their participation and satisfaction with the meeting. Overall we think that the single-session form
of the meeting remains effective, though this does make it more intimidating for younger scientists to ask questions.
The poster sessions work well to showcase the work of SCEC scientists at all levels.
Feedback on the Major SCEC Initiatives
We congratulate SCEC on the success of its major new initiatives. We recognize that these are essential and important
components of the SCEC program, in terms of both scientific scope and funding diversification. These new projects set
the stage for a successful SCEC5. They are also providing SCEC with high political visibility and access.
The CISM initiative will enable improved and more comprehensive physics-based earthquake forecasts to be
developed from evolved models of faulting in California, thus advancing our understanding of faulting hazards. The
AXCESS program will make important computational strides in extending and validating earthquake ground-motion
simulations at higher frequencies (>1 Hz), and facilitating physics-based seismic hazard modeling. The Central
California project holds real promise for both understanding and reducing the uncertainties in ground motion models
that drive seismic hazards at low probabilities. These major SCEC initiatives have transformative potential to increase
our knowledge of earthquake hazards.
Assessment of CEO Advisory Structure and External Evaluation
An initial meeting of the CEO Planning Committee has been convened. It is off to a good start, and helped inform the
direction of CEO for the SCEC5 proposal. It is too soon to make a detailed evaluation of how this structure is working;
in another year we should be better positioned to evaluate its functionality. It would be useful to consider how to
integrate new SCEC products with CEO activities.
We reviewed the CEO Report prepared for SCEC by Michelle Wood. The last few pages of this report were the most
useful. The conclusions and basic recommendations of the Wood report make sense, including the recommendation
to reduce the number of metrics that are tracked. It may be more useful to evaluate in greater depth the effectiveness
of a small number of metrics, rather than gathering many statistics on the accessing of various documents.
SCEC4 Accomplishments
The AC devoted much of its discussions to progress made in SCEC4 in the six fundamental topic areas. We offer the
following observations and suggestions for SCEC as it goes into the final year of SCEC4.
Topic 1: Stress transfer from plate motion to crustal faults: long term slip rates
The imaginative combination of InSAR and GPS spatial and temporal data offers advantages for constraining fault
motions in both the far and near field. A better understanding of locked vs. creeping sections of faults is emerging. The
promise of newly available InSAR products can provide better temporal coverage and orthogonal-look pairs essential
for constraining 3D surface motions in the final year of SCEC4. The discovery of the apparent slip deficit in southern
California from paleoseismic data raises important new scientific questions that will extend into SCEC5.
WELCOME
2016 SCEC Annual Meeting page 16
Topic 2: Stress-mediated fault interactions and earthquake clustering
Particular achievements of note in this topic include the outstanding work on stress and strain modeling. Beyond the
quality, we appreciate the approach to involve all modelers that would like to be involved, and the decision to open the
dataset to everyone. There is still a lot of scientific work to be done as testified by the coordinators and explicitly written
in SCEC5, but the communities have made excellent progress and are on the right track.
Earthquake simulators are the main target of one recent special project of SCEC (CISM). In SCEC4 simulators started
to show their potential, for instance in describing earthquake clustering at different time scales – the short time scale
typical for aftershocks, and a longer time modulation that may potentially explain the clustering observed in
paleoseismic trenches and the so-called open interval conundrum in California seismicity. CISM is a world-leading
initiative.
A retrospective CSEP experiment carried out in New Zealand to forecast the Canterbury sequence shows, for the first
time, that some physics-based models may provide better 1-year forecasts than models based on empirical rules. This
is certainly encouraging for the future SCEC activities in this field.
Overall, the work in this topic is the foundation for operational earthquake forecasting, which is a key direction for SCEC
and for seismic risk mitigation.
Topic 3: Evolution of fault resistance during seismic slip: scale-appropriate laws for rupture
modeling
SCEC4 has made remarkable progress in many diverse areas, ranging from imaging and analysis of fault zone
properties, to modeling the non-linear and plastic contributions to fault slip, and incorporating these elements into
dynamic rupture simulations. The work on dynamic rupture models moves the ground-motion simulation problem from
kinematic models to more fundamental physical behavior of faults and ruptures. The systematic verification of these
models through the dynamic rupture TAG has been ongoing for several years and shows that the models can be used
and get reliable results. Recent studies have addressed the application of the verified dynamic rupture models to
compute ground motions. These studies are mainly sensitivity studies and in many cases show large effects of
parameter variations. What seems to be missing at this point is more comprehensive validation of the dynamic rupture
models against ground motion data. This should be a focus in the final year of SCEC4.
Topic 4: Structure and evolution of fault zones and systems: relation to earthquake physics
Excellent progress has been made in the last 2 years, especially with the flourishing Special Fault Study Areas (SFSA).
Both the San Gorgonio Pass and Ventura SFSAs have been successful at fostering collaborative teams to undertake
and assemble paleoseismic and structural studies of multiple fault strands. These show that multiple strands are
simultaneously active across regions to accommodate slip, at times producing very large earthquakes. SCEC research
in other regions has also demonstrated with geodetic and geologic observations that multiple active strands constitute
broad fault zones that may evolve over the long term.
The Ventura SFSA has added offshore seismic interpretations, including constraints from folding and sedimentation. It
has also incorporated tsunami modelling. The structural models have enabled numerical rupture calculations which
show that throughgoing multiple strand ruptures are possible depending on initial stress and nucleation points.
The SCEC community is on track to successfully complete this SCEC4 component, by considering the probability of
suites of plausible rupture scenarios in the two SFSAs.
Topic 5: Causes and effects of transient deformations
Slow slip events and tectonic tremor SCEC researchers have developed transient detection methodology, but the main
example of such phenomena in southern California remains the 2009 Bombay Beach swarm, which is detectable in
searches for anomalous ETAS behavior of an earthquake swarm, and also in searches for a distinct deformation
transient.
The only triggered tremor identified, which is thought likely to arise from a deformation transient, remains that from the
2002 Denali Alaska earthquake. New ways to search for LFEs continue to be developed, with the hope of finding
deformation transients. This area of investigation appears to have been satisfactorily concluded for SCEC4.
WELCOME
2016 SCEC Annual Meeting page 17
Topic 6: Seismic wave generation and scattering: prediction of strong ground motions
Significant accomplishments in SCEC4 have been made in extending physics-based models of ground motion to higher
frequencies, and in validating simulations to enable their use in engineering applications. These developments hold the
promise of allowing reduced epistemic uncertainty in prediction of ground motions for future large events, which has
tremendous practical significance and cost implications in earthquake engineering.
The validation of the broadband platform is a major step forward in developing physics-based ground motion models
and it is now in a form that can be used for engineering applications. Several recent studies also showed validation of
kinematic ground motion simulation models against empirical data. What has not been addressed is how much do the
ground motion models rely on physics and how much of the performance relies on calibration to empirically recorded
ground motions. For example, if there are parameters in the models that are adjusted to fit the sparse available GM
data, are these models mean-centered? How much better are the constraints than just using the empirical models? For
the last year of SCEC4, it may be useful to try to evaluate how much of the current kinematic models are controlled by
physics and how much is controlled by empirical calibration.
Several other developments in SCEC4 in this topic area are also noteworthy accomplishments. Significant progress
has been made on physics-based high-frequency GM simulations, covering the frequency range from 0 to 10Hz,
including evaluation of the goodness of fit of the simulations; an illustration has been made for the Chino Hills
earthquake. Going even to 5Hz would be a major improvement that might be more achievable.
The importance of inelastic material response effects, both near fault and near surface, in dynamic rupture simulations
has been demonstrated, as applied in particular to CyberShake simulations of large earthquakes on the San Andreas
Fault; expected ground motions are reduced by significant amounts, showing the impact of such effects.
Full wave tomography using earthquake and ambient noise fields can be effective to improve the CVM; these
techniques have been shown to reduce waveform misfits and can ultimately reduce epistemic uncertainty for path
effects in the longer term.
Overall, SCEC4 has been transformative in engaging the earthquake engineering community to realize
practical benefits from the evolution of earthquake process and hazards knowledge. This engagement has
great momentum and provides a compelling rationale for SCEC5: it is expected that in SCEC5 the fruits of this
momentum will be fully realized.
CEO Comments for last year of SCEC4
The expansion of CEO and its increasing level of collaborative activities with IRIS, UNESCO, and engagement of a
broad spectrum of stakeholders, has been a major success for SCEC4. The SCEC/CEO program continues to be a
global flagship for successful CEO activities. The creation of the CEO planning committee is a good step to effectively
target future activities. The recent evaluation and recommendations in the Wood report provide useful guidance for
concluding CEO activities in the final year of SCEC4 and transitioning into SCEC5.
The CEO Director indicated a willingness to identify a list of evaluation research opportunities for graduate students
and early career researchers focusing on existing CEO materials, including scholars in Public Health, Sociology,
Communication, Education, Marketing, etc. and would be a good conclusion to SCEC4 CEO efforts.
In conclusion, the AC continues to be deeply impressed with the amazing quality of science and collaboration, not to
mention the boundless energy, that the SCEC community brings to the task of understanding earthquake hazards in
Southern California.
WELCOME
2016 SCEC Annual Meeting page 18
Communication, Education, and Outreach Highlights
Mark Benthien, SCEC Associate Director for CEO
Overview
SCEC’s Communication, Education, and Outreach (CEO) program facilitates learning, teaching, and application of
earthquake research. In addition, SCEC/CEO has a global public safety role in line with the third element of SCEC’s
mission: “Communicate understanding of earthquake phenomena to end-users and society at large as useful
knowledge for reducing earthquake risk and improving community resilience.” The theme of the CEO program during
SCEC4 has been Creating an Earthquake and Tsunami Resilient California. Our geographic reach has expanded far
beyond the Golden State via partnerships across the country and worldwide. The goal is to prepare people for making
decisions about how to respond appropriately to changing seismic hazards, including tsunami warnings and new
technologies such as operational earthquake forecasting and earthquake early warning.
SCEC/CEO has been very successful in leveraging its base funding with additional support. For example, since 2010,
FEMA has provided SCEC nearly $1.5 million to coordinate the Earthquake Country Alliance in California (at the request
of the California Office of Emergency Services, CalOES) and for national ShakeOut coordination. ShakeOut regions in
the U.S. and internationally have also provided funding, and the California Earthquake Authority (CEA) has spent
several million dollars on advertising that features ShakeOut promotions each year. SCEC’s intern programs have been
supported with more than $1.3 million in additional funding from several NSF programs and a private donor, and NASA
supports SCEC’s “Vital Signs of the Planet” teacher development program (via JPL) as part of the NASA InSight
mission. NOAA (via CalOES) now provides funding to SCEC for developing the TsunamiZone.org website.
Evaluation of the CEO program is conducted each
year by SCEC’s external Advisory Council, via annual
reporting of milestones and metrics to funding
agencies, as part of individual activities (post-
ShakeOut surveys, teacher workshop evaluations,
post-internship discussions, etc.), and as part of
proposal reviews. In Spring 2015 a new “CEO
Planning Committee” comprising members of the
SCEC Advisory Council as well as SCEC community
stakeholders was established to help guide and
support SCEC/CEO activities and partnerships, which
have significantly expanded during SCEC4. In
addition, an experienced program evaluator has
reviewed the CEO program overall including its
evaluation structures. Analyses for each CEO area
were provided along with recommendations for how to
expand and improve evaluation, including a new
comprehensive logic model to tie all CEO activities to
a set of long term intended outcomes. The results
indicate that the SCEC/CEO program plays an
important role in earthquake education and
preparedness (Box 1), and the evaluation’s
recommendations have influenced the CEO program
plan for SCEC5.
WELCOME
2016 SCEC Annual Meeting page 19
Major Activities and Results
Global network of Great ShakeOut Earthquake Drills,
and related campaigns
Great ShakeOut Earthquake Drills began in southern California in
2008, based on the USGS-led “ShakeOut Scenario” for a large
(M7.8) San Andreas earthquake. ShakeOut communicates
scientific and preparedness information with the mission to
motivate everyone, everywhere to practice earthquake safety
(“Drop, Cover, and Hold On”), and to promote resiliency through
preparedness and mitigation.
Working with a small committee of Earthquake Country Alliance
(ECA) leaders, SCEC created an online registration system and
resource site (ShakeOut.org) where more than 5.4 million
southern Californians were registered to participate in 2008,
building to 10.5 million people statewide in 2015 (Box 2). While K-
12 and college students and staff comprise the largest number of
participants, ShakeOut has also recruited businesses, non-profits,
government agencies, neighborhoods groups, and individuals.
In addition to leading the California ShakeOut, SCEC manages a
network of ShakeOut Regions worldwide, and hosts the website
for each of their drills (except Japan). As of 2015, 26 Official
ShakeOut Regions span 51 states and U.S. territories, three
Canadian provinces, New Zealand, Southern Italy (U.S. Naval
bases), and Japan. People and organizations in any other state or
country can also register to be counted in the overall global total.
More than 26.5 million people were registered in 2014. SCEC’s
Associate Director for CEO, Mark Benthien, was recognized as a
“White House Champion of Change” for leading these efforts, and
FEMA has based its national “America’s PrepareAthon!” multi-
hazard campaign on ShakeOut to assess preparedness activities
for other hazards (and contracts with SCEC for consultation).
ShakeOut has become a global infrastructure for providing earthquake information to the public and involving them in
community resiliency. New countries are being actively recruited to join the ShakeOut movement, which serves to
coordinate earthquake messaging internationally. Participants receive monthly ShakeOut newsletters and more
frequent content via social media. Millions more learn about ShakeOut via broad news media coverage that encourages
dialogue about earthquake preparedness. Surveys of ShakeOut participants show increased levels of mitigation and
planning, and encouragement of peers to participate and get better prepared. In the near future, ShakeOut will be
utilized for educating Californians about Earthquake Early Warning, with yearly tests to be held on ShakeOut day.
As a result of its leadership of ShakeOut, SCEC now also receives NOAA funding provided through the California Office
of Emergency Services to create and manage TsunamiZone.org. This international site adapts the ShakeOut
registration system to assess participation in Tsunami activities, whether as part of their ShakeOut activities or during
local tsunami preparedness weeks or months. Primary participation in 2016 included California, Oregon, Washington,
Hawaii, and more than 20 countries of the Caribbean.
Extensive collection of public education and preparedness resources and activities
Partnerships. The Earthquake Country Alliance (ECA) was created in southern California by SCEC with many partners
in 2003 and is now a statewide coalition with similar groups in the Bay Area and North Coast. ECA’s sector-based
committees develop consistent messaging and resources distributed via activities led by each regional alliance. SCEC’s
Associate Director for CEO Mark Benthien is ECA’s Executive Director. In 2012 ECA received FEMA’s “Awareness to
Action” award and also the “Overall National Award in Excellence” at the National Earthquake Conference, both for its
WELCOME
2016 SCEC Annual Meeting page 20
creation of ShakeOut and other activities. In 2014 ECA was given an award from the American Red Cross for
“Excellence in Disaster Preparedness”.
SCEC also coordinates the Earthquake and Tsunami Education and
Public Information Center (EPIcenter) Network of more than 60
museums, science centers, and libraries, some of which host SCEC-
developed exhibits and programming. SCEC has also established
“EPIcenters” in other states (Oregon, Alaska, Maine, and others), and is
working with the Central United States Earthquake Consortium (CUSEC)
to create a Central U.S. EPIcenter network. CUSEC also manages the
Central U.S. and Southeast multi-state ShakeOut regions.
Resources. In addition to the SCEC.org website, SCEC develops and
maintains all ECA websites—EarthquakeCountry.org,
DropCoverHoldOn.org, and Terremotos.org (Figure 1)—and the global
websites ShakeOut.org and TsunamiZone.org. These sites allow SCEC
to promote consensus-based messaging. In 2014 a “Northridge
Earthquake 20th Anniversary Virtual Exhibit” was added to the ECA site,
including “Northridge Near You” animations created by SCEC UseIT
interns. Similar animations plus related graphics were made for the Loma
Prieta 25th anniversary, again by UseIT interns. In addition, SCEC’s
social media presence has greatly expanded since 2013 with active Twitter,
Facebook, YouTube, and other accounts for SCEC, ECA, ShakeOut, and
TsunamiZone distributing SCEC and ECA messaging globally.
SCEC’s Putting Down Roots in Earthquake Country handbook (Figure 2) has
provided earthquake science and preparedness information to southern
Californians since 1995. The 2004 update introduced the Seven Steps to
Earthquake Safety, the main organizing structure for SCEC, ECA, and CEA
preparedness messaging. Related versions are now available in multiple
languages, for businesses, and for the San Francisco Bay Area, California’s
North Coast, Nevada, Oregon, Utah, the Central U.S., and Idaho. In 2014 the
California Earthquake Authority, California Office of Emergency Services, and
ECA created a simpler booklet, Staying Safe Where the Earth Shakes, with
customized versions for 10 regions of the state and multiple language editions
(Spanish and Chinese to start).
Additional resources developed by SCEC and ECA Associates during SCEC4
include earthquake safety materials and ShakeOut guidelines for seniors and people with disabilities, higher education,
government agencies, businesses, and healthcare facilities. SCEC developed (with support of USC students) the
“Earthquake Safety Video Series” at www.youtube.com/greatshakeout, a growing collection of “airplane safety” type
short videos demonstrating what to do in various situations during earthquakes. SCEC and Outreach Process Partners,
with FEMA funding designated by CalOES, developed a multi-location earthquake safety poster which will be available
for free from the FEMA Warehouse (Figure 3). SCEC and the ECA Seniors and People with Disbility Committee also
made a new graphic for people with mobility limitations, showing what to do if someone uses a wheelchair, walker, or
cane (see www.EarthquakeCountry.org/disability).
WELCOME
2016 SCEC Annual Meeting page 21
Activities. SCEC and its partners coordinate a broad range of public education and collaboration activities. Each ECA
Regional Alliance holds several workshops each year featuring guest speakers and outreach planning. The ECA
Speakers Bureau holds monthly trainings at SCEC, and members speak to community groups, businesses, and other
organizations, and staff tables at preparedness fairs. ECA sector-based committees presented a series of ECA
webinars in 2015 to nationwide audiences. EPIcenter locations hold public lectures and host day-of ShakeOut events.
SCEC staff also participate in local government, nonprofit, and business meetings throughout the year, and host foreign
groups interested in our best practices.
News media coordination. SCEC continues to develop new procedures for post-earthquake media coordination,
because the breadth of SCEC’s research, including its information technology programs and the development of time-
dependent earthquake forecasting, is increasing the need for expanded media relations. New strategies and
technologies are being developed to meet these demands, such as the use of a media relations service for identifying
and connecting with reporters nationwide and then tracking resulting news coverage (used for both SCEC and
ShakeOut media coordination). SCEC also partners with USGS, IRIS, Caltech, and other partners to offer programs
that educate the media on how to report earthquake science. Examples include a media training workshop at Caltech
and a press conference at USC as part of the 20th Anniversary of the Northridge Earthquake in January 2014. In 2015
SCEC coordinated with USGS, CalOES, FEMA and other partners to address issues with the movie San Andreas,
including numerous interviews and resources organized by SCEC including “fact or fiction” analysis (see
www.earthquakecountry.org/sanandreas). The response also included extensive social media engagement, for which
SCEC created the “Seven Steps to Earthquake MOVIE Safety”, a parody of our standard Seven Steps messaging
(www.earthquakecountry.org/moviesafety).
Broad range of K-14 educator partnerships, programs, and resources
Workshop Partnerships. SCEC is an active participant in the science education community including local and
national organizations such as the California Science Teachers Association (CSTA). In 2011 and 2013 SCEC
participated in the planning committee for CSTA’s Annual Conference and sponsored a 2013 keynote talk by SCEC
intern alumnus Emmett McQuinn. Since 2009, SCEC has hosted earthquake-oriented field trips and workshops for
WELCOME
2016 SCEC Annual Meeting page 22
more than 150 teachers. In addition, SCEC and the California Geological Survey co-host a booth at CSTA meetings
that draw ~2000 attendees each year.
SCEC also collaborates in this area with other Earth science organizations such as its active role in EarthScope’s
workshops for park and museum interpreters since 2008. SCEC has participated in all four of the Cascadia EarthScope
Earthquake and Tsunami Education Program (CEETEP) workshops in the Pacific Northwest, which have served over
100 educators, emergency managers and park interpreters. SCEC is co-hosting the final workshop in Arcata, California,
in fall 2015. SCEC’s EarthScope partners have found that the ShakeOut is an important event that helps promote their
program (and vice versa).
InSight Vital Signs of the Planet (VSP) Professional
Development Program. SCEC has a lead role in the
education program for InSight (Interior Exploration using
Seismic Investigations, Geodesy, and Heat Transport), a
NASA mission that will place a geophysical lander on Mars
(the mission was delayed but is again on track). SCEC
developed the VSP program, a research experience and
curriculum development program for K-12 teachers that
expands on a collaboration between SCEC and the Cal State
San Bernardino/EarthScope RET program led by Dr. Sally
McGill. Since 2013 VSP has provided 30 educator fellows
(and select students) experiences in scientific inquiry
including a 5-day field experience using GPS to monitor
tectonic deformation in Southern California (with instruments
provided by UNAVCO) (Figure 4). Participants then develop
and test lesson plans and convene a workshop held during the
SCEC Annual Meeting. The program is complete for now,
awaiting potential resumption with the re-start of the NASA
mission. In the mean time, the relationships developed in the
program have been proposed as part of a national NSF-
funded diversity pathway project of which Dr. McGill is a Co-
PI and SCEC will participate.
Quake Catcher Network (QCN). SCEC has expanded QCN
with installations of low cost seismometers at over 26
EPIcenter museum locations in California and Oregon, at
more than 100 schools in each west coast state including
Alaska, and in more than 30 schools and museums in the
Central U.S., including most recently installations in Oklahoma
(Figure 5). The goal is to establish several K-12 sensor
stations around a local museum hub as a means to build long-
term educational partnerships around the ShakeOut, citizen science, and enrich K-12 STEM curriculum. In 2015 a new
partnership was established between SCEC, IRIS, and USGS to continue the expansion and development of QCN
worldwide, with SCEC hosting the QCN servers at USC, and IRIS managing the website.
Plate Tectonics Kit. This teaching tool created and distributed by SCEC was developed to make plate tectonics
activities more accessible for science educators and their students. SCEC developed a user-friendly version of the This
Dynamic Planet puzzle map, which is used to teach about plate tectonics. Educators often suggested that lines showing
the location of plate boundary on the back of the maps would make it easier for them to correctly cut the map, so SCEC
designed a new, two-sided map.
Well-established undergraduate research experiences
The SCEC Experiential Learning and Career Advancement (ELCA) program enhances the competency and diversity
of the STEM workforce by (1) engaging students in research experiences at each stage of their academic careers and
(2) providing leadership opportunities to students and early career scientists that engage them in the SCEC community.
ELCA manages two undergraduate internship programs that involve over 30 students each summer:
WELCOME
2016 SCEC Annual Meeting page 23
● The Summer Undergraduate Research Experience (SURE) program places undergraduate students with
SCEC scientists around the country. More than 270 interns have participated since 1994. Projects have
spanned all areas of earthquake science, engineering, and education.
● The Undergraduate Studies in Earthquake Information Technology (UseIT) program brings together
students from across the country to an NSF Research Experience for Undergraduates Site at USC. The eight-
week program develops computer science skills while teaching the critical importance of collaboration for
successful learning, scientific research and product development. Since 2002, 293 internships have been
supported. USEIT interns tackle a scientific “Grand Challenge” each year that entails developing software and
resources for use by earthquake scientists or outreach professionals. The 2016 Grand Challenge was to run
physics-based earthquake simulators on a high-performance computer to generate long catalogs of California
seismic activity, develop 10-year forecasts of large earthquakes (M ≥ 7) on the southern San Andreas fault
system based on these catalogs and compare these to UCERF3, and then identify threatening multi-event
scenarios within the catalog and illustrate them with hazard and risk maps.
Since 2002, over 1600 eligible applications have been submitted to the SCEC internship programs (at
www.scec.org/internships). Since 2010, underrepresented minority interns averaged 36.4% of each year’s class, with
a high of 43% in 2014 (Figure 6). Women represented an average of 48% of interns, with a high of 57% in 2014. First-
generation college attendees have averaged 31% of each class.
Much of the success in increasing diversity has come from
increased efforts to recruit students from other states and also
from community colleges, making the internship programs an
educational resource that is available to a broader range of
students.
Past interns report that their internship made lasting impacts on
their course of study and career plans, often influencing
students to pursue or continue to pursue earthquake science
degrees and careers. By observing and participating in the daily
activities of earth science research, interns reported having an
increased knowledge about working in research and education,
which coupled with networking at the SCEC annual meeting,
gave them the inspiration and confidence to pursue earth
science and career options within the field.
Knowledge Implementation
SCEC produces a large body of knowledge about the seismic hazard in California that enhances seismic hazard maps,
datasets, and models used in building codes and engineering risk assessments. The Earthquake Engineering
Implementation Interface, led by Jack Baker and Jacobo Bielak, provides the organizational structure for creating and
maintaining collaborations with research engineers to ensure SCEC’s research activities are aligned with their needs.
These activities include rupture-to-rafters simulations of building response as well as the end-to-end analysis of large-
scale, distributed risk (e.g., ShakeOut-type scenarios). Analysis of the performance of very tall buildings in Los Angeles
using end-to-end simulation remains a continuing task that requires collaboration with both research and practicing
engineers through PEER and other organizations. An important Technical Activity Group in SCEC4 has been the
Ground Motion Simulation Validation (GMSV) group, led by Nico Luco and Sanaz Rezaeian, which is developing
procedures for the validation of numerical earthquake simulations that are consistent with earthquake engineering
practice.
The Implementation Interface also develops mechanisms for interacting with technical audiences that make decisions
based on an understanding of earthquake hazards and risk, including practicing engineers, geotechnical consultants,
building officials, emergency managers, financial institutions, and insurers. An example is the annual SEAOSC
Buildings at Risk Summits, which SCEC has co-organized since 2011 in both Los Angeles and San Francisco (with
SEAONC). The 2015 conference was titled “Strengthening our Cities.” In 2014 SCEC/ECA also helped create the
“Earthquake 2014 Business Preparedness Summit” with FLASH, Safe-T-Proof, Simpson Strongtie, and several other
partners, which launched a new QuakeSmart recognition program for businesses that demonstrate mitigation they have
implemented. These summits have since been offered in several locations nationwide.
WELCOME
2016 SCEC Annual Meeting page 24
Research Accomplishments
Gregory C. Beroza, SCEC Science Planning Committee Chair
Judith S. Chester, SCEC Science Planning Committee Vice-Chair
Overview
The SCEC4 theme of “Tracking Earthquake Cascades,” expressed the fact that seismic hazard varies strongly with
time. Understanding and quantifying this time-dependence across all time scales of scientific and societal interest
motivated the SCEC4 core research program. The SCEC4 science plan resolved the challenge of tracking earthquake
cascades into six fundamental questions of earthquake physics:
● Stress transfer from plate motion to crustal faults: long-term fault slip rates
● Stress-modulated fault interactions and earthquake clustering: evaluation of mechanisms
● Evolution of fault resistance during seismic slip: scale-appropriate laws for rupture modeling
● Structure and evolution of fault zones and systems: relation to earthquake physics
● Causes and effects of transient deformations: slow slip events and tectonic tremor
● Seismic wave generation and scattering: prediction of strong ground motions
These questions formed the interdisciplinary framework to develop a better understanding of time-dependent hazard.
In what follows we report on the accomplishments as organized into the SCEC4 disciplinary and interdisciplinary focus
groups.
1. Seismology
The Seismology Group gathers data on the range of seismic phenomena observed in southern California and integrates
these data into seismotectonic interpretations as well as physics-based models of fault slip. Resources include the
Southern California Earthquake Data Center (SCEDC) that provides extensive data on Southern California earthquakes
as well as crustal and fault structure, the network of SCEC funded borehole instruments that record high quality
reference ground motions, and the pool of portable instruments that is operated in support of targeted deployments or
aftershock response.
This past year’s accomplishments include:
● Continued archiving and distribution of seismic waveforms from 507 southern California seismic stations.
Updated catalogs of relocated earthquakes and focal mechanisms are available.
● The finding that induced sequences near active geothermal production tend to be short in duration but have
high aftershock productivity.
● Fluid pathways along known faults can result in events being induced at large (10 km) distances from injection
wells.
● Ambient seismic field Green’s functions enable prediction of ground motions for future moderate-to-large
earthquakes.
● Improvements to the Community Velocity Model are being enabled by comprehensive analysis of dense
datasets along the San Jacinto Fault Zone.
● Detailed analysis of two earthquake sequences, 2015 Fillmore and 2016 Mw5.2 Borrego Springs
1.1 Data Center, Data Products, and Instrumentation
The Southern California Earthquake Data Center (SCEDC) archives continuous and triggered data from 507 Southern
California Seismic Network (SCSN) stations. These archives include data from approximately 16,000 earthquakes each
year, on average. In addition to raw data, SCEDC generates a myriad of products that are vital to emergency response
and earthquake research (Clayton and Yu, 2016). Community data products of relocated seismicity, improved focal
mechanisms, and waveform spectra are widely used by SCEC and other researchers for a variety of studies and
continue to be a key resource. A new version of the relocated catalog (Hauksson, Yang, and Shearer) provides precise
locations for over 582,000 events that occur from 1981 through 2015 (Hauksson and Shearer, 2016). An updated focal
mechanism catalog provides refined focal mechanisms for events that occurred through 2014. This catalog includes
improvements to the processing by using the latest relocated catalog for event locations, making use of previously
downloaded data, and improved handling of metadata. The SCEC community also benefits from instrumentation and
WELCOME
2016 SCEC Annual Meeting page 25
data made available through the borehole program and portable broadband instrument center (PBIC) (Steidl, 2016a).
The borehole program leverages resources from several organizations to provide this valuable dataset. Data are
integrated into the SCEDC for easy access by the seismological community (Steidl, 2016b). PBIC provides a local
resource of instrumentation for rapid installation following a major earthquake in southern California and, between major
events, it serves as an instrument pool for individual PI-driven research.
1.2. Induced Seismicity
Efforts to identify and characterize induced seismicity are being pushed forward by a number of studies supported by
SCEC. One area of focus is the geothermal fields near the Salton Sea where comprehensive and long-term seismic
data sets exist. New detection techniques, such as matched filter detection, are being used to show that microseismicity
increases following distant mainshocks and suggest the area of active geothermal production is critically stressed
(Casey et al., 2015; Peng, 2016). A complementary study uses data from a local borehole network operated by
CalEnergy to obtain high-resolution earthquake locations in the Salton Sea Geothermal field. These catalogs are then
used to probe the characteristics of induced events (Cheng and Chen, 2015; Chen, X. 2016). Seismicity near active
geothermal fields is found to occur in rapid bursts that are short in duration but with high productivity of small
aftershocks, i.e. high b-value. In contrast, regions farther from active energy production events have more typical (i.e.
lower) b-value statistics. Another area of study is understanding the conditions that can lead to events being induced
by wastewater injection. A SCEC-funded
study drew on the availability of detailed fault
maps to create a flow model of pressure
perturbations near an active injection well. The
model allowed them to investigate the
possible fluid pathways that linked an injection
well with seismicity occurring on a major fault
structure located 10 km away (Figure 1.1)
(Brodsky and Goebel, 2016). SCEC
researchers have also leveraged academia-
industry partnerships to gain access to unique
datasets of seismic data during hydraulic
fracturing stimulation taking place within 8 km
of the San Andreas fault. Using new detection
techniques over 12,000 events were identified
within a 7.5 hour injection period (Chen, J.,
2016). The events show a deviation from
Gutenberg-Richter statistics similar to what
has been observed in previous studies of
induced sequences.
1.3. 3D Crustal Structure and Ground
Motion
Mapping the spatial variability in crustal velocity, particularly close to fault zones and in sedimentary basins, enables
improvements in understanding and predicting ground motions. The use of ambient noise tomography in constructing
Green’s functions has proven to be an effective method for predicting ground motions for large events, including in
sedimentary basins. Virtual M7 earthquakes based on the San Jacinto Fault and the Itoigawa-Shizuoka Tectonic Line
are being used to predict ground motions in the Los Angeles and Tokyo metropolitan areas using multi-year stacks
ambient seismic field Green’s functions (Denolle and Vernon, 2016; Boue et al., 2016) (Figure 1.2). Similarly, dense-
array recordings of the ambient noise field in Long Beach are being used to develop random-field representations of a
3D P wave velocity model (Beroza and Nakata, 2016). The model suggests that there are strong small-scale length
variations and are validated by comparisons to code of nearby earthquakes. High resolution imaging of 3D crustal
structure has been undertaken along the San Jacinto Fault Zone using multiple datasets and approaches including P
and S arrivals, ambient noise dispersion, joint inversion, and identification of fault head waves (Thurber and Ben-Zion,
2016). The results will be incorporated into the Community Velocities models with the goal of improving ground motion
predictions.
WELCOME
2016 SCEC Annual Meeting page 26
1.4 The 2015 Fillmore Earthquake Swarm
The 2015 Fillmore swarm occurred about 6 km west of the City of Fillmore in Ventura, California, and was located
beneath the eastern part of the actively subsiding Ventura basin at depths from 11.8 km to 13.8 km, similar to two
previous swarms in the area (Figure 1.3). Template-matching event detection showed that it started on the 5th of July
2015 at 2:21 UTC with a ~M1.0 earthquake.
Figure 1.3. A perspective view looking west of the Ventura basin and Fillmore swarm. The different color surfaces are major late Quaternary faults represented in the SCEC Community Fault Model (CFM (Plesch et al., 2007). The coast line is shown as a white jagged line. Hypocenters for the 2015 Fillmore swarm are highlighted by magnitude, and occur along the down-dip extent of the Simi-Santa Rosa fault zone Nicholson et al., (2014).
The swarm exhibited unusual episodic
spatial and temporal migrations, and
diversity in nodal planes of focal
mechanisms as compared to the simple
hypocenter defined plane. It was also
noteworthy because it consisted of >1,400
events of M≥0.0, with M2.8 being the
largest event. We suggest that fluids
released by metamorphic dehydration processes, migration of fluids along a detachment zone, and cascading asperity
failures caused this prolific earthquake swarm. Other mechanisms such as simple mainshock-aftershock stress
triggering or a regional aseismic creep event are less likely. Dilatant strengthening may be a mechanism that causes
the temporal decay of the swarm as pore pressure drop increased the effective normal stress, and counteracted the
instability driving the swarm (Hauksson et al., 2016).
1.5 The 2016 Borrego Springs Earthquake Sequence
The trifurcation area of the San Jacinto fault zone has produced more than 10% of all earthquakes in southern California
since 2000, including the June 2016 Mw 5.2 Borrego Springs earthquake (Figure 1.4).
WELCOME
2016 SCEC Annual Meeting page 27
Figure 1.4. Map of the trifurcation area of the San Jacinto fault zone. Historical seismicity is denoted by black dots (20). Events with M > 4.0 that have occurred in this area since 2001 are indicated by blue stars and focal mechanisms. Aftershocks of the 2016 Borrego Springs earthquake are colored by depth. Red stars and focal mechanisms indicate aftershocks with M > 3.0. Stations used are denoted by green triangles. Figure insets contain a close-up of the detected aftershocks, which delineate numerous strike-slip and normal faults in conjugate orientations, and the b-value plot for the ~25,000 aftershocks. For the first time ever, we can report magnitude of completeness of M-0.5.
Here the fault splits into three sub-parallel strands and is associated with broad Vp/Vs anomalies. We synthesize spatial
properties and rates of early aftershocks of this event and historical background seismicity. A template matching
technique is used to detect and locate more than ~25,000 early aftershocks, which together with high-resolution
regional seismicity, illuminate highly complex active fault structures. The hypocenters form dipping lineations of
seismicity both along-strike and nearly normal to the main fault, and are composed of interlaced strike-slip and normal
faults. The three main strike-slip faults change dip with depth and become listric, by transitioning to a dip of ~70 degrees
near 10 km depth. Events with M > 4 occurred on the three main faults, while most of the small magnitude events
occurred in a damage zone several kilometers wide at seismogenic depths. This spatial separation of events in different
magnitude ranges manifests in a bimodal frequency-magnitude distribution. The anomalously high seismicity rate in
the trifurcation area is likely a reflection of the extraordinary geometric complexity. The results further provide important
insights into the physics of faulting at depth, near the brittle-ductile transition (Ross et al., submitted, 2016).
1.6. Select Publications
● Beroza, G. and N. Nakata, 2016, Characterizing spatial variability of ground motion using very dense arrays,
SCEC Annual Report Award #15213.
● Boué, P., M. Denolle, N. Hirata, S. Nakagawa, and G. C. Beroza, 2016, “Beyond Basin Resonance:
Characterizing Wave Propagation Using a Dense Array and the Ambient Seismic Field”, Beyond Basin
Resonance: Characterizing Wave Propagation Using a Dense Array and the Ambient Seismic Field, Geophys.
J. Int., doi: 10.1093/gji/ggw205.
● Brodsky, E. E. and T. H. W. Goebel, 2016, Assessing fault zone structure and permeability in regions of active
faulting and fluid injection: Can fault maps and structure help evaluate induced seismicity in southern
California, SCEC Annual Report: Award #15168.
● Casey, B., X. Meng, D. Yao, X. Chen and Z. Peng, 2015, Systematic detection of remotely triggered seismicity
in Salton Sea with a waveform matching method, abstract submitted to the annual Southern California
Earthquake Center meeting, Palm Springs, CA.
WELCOME
2016 SCEC Annual Meeting page 28
● Chen, J., 2016, Characterization of induced micro-seismicity associating with one hydraulic fracturing
experiment near the San Andreas fault, central California, SCEC Annual Report Award #15137.
● Chen, X., 2016, Effect of geothermal operations on earthquake source spectra, SCEC Annual Report Award
#15026.
● Cheng, Y., and X. Chen, 2015, Effect of geothermal operations on seismic characteristics in the Salton Sea
geothermal field, SEG Tech. Progr. Expand. Abstr., 5063–5068, doi:10.1190/segam2015-5882527.1.
● Clayton, R. W., and E. Yu, 2016, Southern California Earthquake Data Center (SCEDC) 2015 operations,
SCEC Annual Report Award #15061.
● Denolle, M. and F. Vernon, Response to virtual earthquakes on the San Jacinto Fault and on the Itoigawa-
Shizuoka Fault, SCEC Annual Report Award #15036.
● Goebel, T. H. W., S. M. Hosseini, F. Cappa, E. Hauksson, J.-P. Ampuero, F. Aminzadeh & J. B. Saleeby,
2016, “Wastewater disposal and earthquake swarm activity at the southern end of the Central Valley,
California”, Geophys. Res. Letts., 43, 1-8, doi:10.1002/2015GL066948.
● Hauksson, E., J. Andrews, A. Plesch, J. H. Shaw, and D. R. Shelly, The 2015 Earthquake Swarm Near
Fillmore, California: Possible Dehydration Event Near the Bottom of the Over-Pressurized Ventura Basin, in
press, Seismo., Res. Letters, v87, nr4, p807-815, doi: 10.1785/0220160020, 2016.
● Peng, Z., 2016, Comprehensive detection of injection-induced seismicity in the Salton Sea Geothermal field,
SCEC Annual Report Award #15077.
● Ross, Z. E., E. Hauksson, Y. Ben-Zion, Orthogonal faults and bimodal seismicity rates from trifurcation of the
San Jacinto fault, submitted to Science Advances, Aug. 2016.
● Shearer, P. M. and E. Hauksson, SCEC community data products of relocated seismicity, improved focal
mechanisms, and waveform spectra for resolving fine-scale fault structures and state of stress in southern
California, SCEC Annual Report Award #15017.
● Steidl, J, 2016a, SCEC portable broadband instrument center (PBIC), SCEC Annual Report Award #15185.
● Steidl, J., 2016b, SCEC borehole instrumentation program, SCEC Annual Report Award #15172.
● Thurber, C. and Y. Ben-Zion, 2016, Joint inversion of direct P and S waves, head waves and noise dispersion
data for the San Jacinto fault region, SCEC Annual Report Award #15112.
WELCOME
2016 SCEC Annual Meeting page 29
2. Tectonic Geodesy
A main focus of the SCEC Tectonic Geodesy (TG) activities has been the continuing development of the Community
Geodetic Model (CGM), a crustal motion model that will ultimately consist of velocities and deformation time series for
southern California that leverages the complementary nature of Global Positioning System (GPS) and Interferometric
Synthetic Aperture Radar (InSAR) observations. This project is coordinated by TG Leaders Sandwell and Funning, and
past TG Leaders Murray and Lohman through in-person workshops (most recently in January 2016), and online
interactions. In addition to CGM-focused work, group members continue to work on other topics of interest, including
studies of fault creep, transient deformation (both tectonic and non-tectonic) and fault slip rates.
This past year’s accomplishments include:
● A rigorous comparison of continuous GPS time series from multiple independent processing centers.
● A comprehensive reprocessing of the entire southern California campaign GPS archive including data
collected up to 2014.
● Development of a consensus CGM horizontal velocity grid.
● Methods to combine InSAR and GPS to solve for the full 3D deformation field at high resolution.
● Ongoing efforts to densify geodetic measurements in regions of interest through campaign GPS
measurements.
● Methodological improvements in the study of vertical deformation, its relationship to hydrological storage and
temporal correlation with seismicity.
● The use of deviations from the expected shear strain rate along the San Jacinto fault to infer the presence of
a zone of reduced elastic strength.
2.1 Community Geodetic Model
After several years of methodological
development and exploration of the
data, some significant progress has
been made in the last year towards the
goals of the CGM project. Much of the
research summarized here was
presented and discussed by CGM
participants at a SCEC-funded
workshop held in Pomona, CA on
January 28–29, 2016.
Herring and Floyd (annual report) have
made headway in their efforts to
compare, continuous GPS time series
produced by five different processing
centers, with the ultimate goal of
combining these into a single product. If
results are aligned to a common
reference frame without estimating a
scale, then the resultant time series
show a high degree of similarity. When
offsets, seasonal and postseismic
transients are removed, secular rates of deformation agree within 0.15 mm/yr for the horizontal components and 0.5
mm/yr for the vertical (Figure 2.1). These results suggest that robust velocity and time series products can be produced
by averaging the different groups’ time series. Shen (annual report) in a complementary effort, has reprocessed the full
archive of campaign GPS data acquired between 2002 and 2014 in Southern California, including new data from
multiple groups (U. Arizona, CSU San Bernardino, UC Riverside, Scripps), resulting in improved coverage in the
Eastern Transverse Ranges in particular.
The reprocessed campaign GPS data are being used to develop a consensus CGM horizontal velocity grid at 0.01˚
resolution (Figure 2.2, Sandwell et al., 2016 annual meeting). The approach is to first assemble 15 available (mostly
WELCOME
2016 SCEC Annual Meeting page 30
published) velocity and strain rate models for the SCEC region. The 15 strain rate models were evaluated in terms of
roughness, cross correlation, seismicity rate, and SHmax to select a subset of 10 usable models. All the models are
based on slightly different geodetic data and use a variety of reference frames so they were regridded at a 0.01˚ grid
spacing using the common campaign GPS velocity vectors (i.e., remove/interpolate/restore method). The 10 velocity
models were averaged (Figure 2.2) and their standard deviation was also computed (Figure 2.3). Standard deviations
are generally small (< 0.5 mm/yr) in areas of good GPS coverage; areas of large standard deviation illustrate where
InSAR velocities will contribute most. This uniform velocity grid is a first step in the development of the full 3-D time
dependent CGM. As new GPS and InSAR
data become available this SCEC
community model will continue to evolve.
The full compilation of the GPS velocity
data, the contributed models, and the
consensus products will be available on
the SCEC website.
Figure 2.2 CGM Horizontal Velocity Grad, V1 is
the mean velocity of 10 contributed models.
Colors show velocity magnitude and arrows
show direction. Black lines are faults from
California and Nevada.
Figure 2.3 Standard deviation of the velocity of
the GGM velocity. Note the deviations are small
in the locations of the GPS constraints (black
dots), especially in areas where the velocity
model is smooth. Dotted gray lines are faults
from California and Nevada.
Advances were also made in the
combination of InSAR and GPS into
integrated data products. Liu and
collaborators (Annual Meeting abstract;
Liu et al., 2015) have demonstrated a
method to estimate a 3D velocity field by
first interpolating horizontal GPS vectors
onto a common grid with InSAR data using
an algorithm that takes into account the
GPS station density and configuration, and
then inverting for the vertical component
from the InSAR using a weighted least squares inversion. Lohman et al. (2015 annual project report) have continued
their analysis of methods for integrating InSAR and GPS time series and in particular explored methods for combining
3-D GPS vectors with 1-D line-of-sight (LOS) InSAR data. Schmidt et al. (2015 annual project report) present a method
for recovering the time-dependent deformation that enhances the signal in regions of low coherence. Their
InSAR time-series method incorporates pixel coherence into the covariance matrix, which performs better
than the traditional Small Baseline Subset (SBAS) method. An example of the improvement in LOS velocity
estimation in the Coacella Valley is provided in Figure 2.4.
WELCOME
2016 SCEC Annual Meeting page 31
Figure 2.4. Mean LOS velocity constructed by two different methods. A) Coherence-based SBAS; B) SBAS with uniform weights. The data set is from ERS-1 and ERS-2 (1992-2001). Red is ground moving away from satellite and blue is ground moving toward satellite. Major fault segments are marked byblack dots. The standard deviation of the mean velocity is denoted in the bottom of each figure.
The TG group have continued to support GPS data collection, both to increase density of observation in critical areas
(e.g. the San Gorgonio Pass and Ventura SFSAs), and
also to update site positions, both in order to improve
uncertainties in velocity for use in the CGM and also in an
ongoing readiness effort for earthquake response. In the
past year, McGill, Bennett and Spinler (annual report;
McGill et al., 2015), led a project in the San Bernardino
mountains and Joshua Tree National Park where
undergraduate interns, high school students and teachers
were involved in GPS data collection at over 35 sites
(Figure 2.5). In a similar effort, Funning (annual report)
and an undergraduate intern collected GPS data and
processed updated site positions at 37 sites, mostly
located in the San Jacinto valley area. Data from both
projects have been uploaded to the campaign GPS
archive at UNAVCO. Finally, Sandwell, Gonzalez-Ortega
and colleagues collected high-resolution campaign GPS
spanning the Imperial and Cerro Prieto faults in northern
Baja California, leading to better constraints on the slip
rates of those two plate-boundary faults.
2.2. Transient deformation
In the past year Borsa and Becker (project report; CGM workshop presentation) focused their efforts on improving
estimates of GPS vertical deformation time series through estimating, and correcting for, the effects of loading from
terrestrial hydrology. They developed a method for estimating the effect of terrestrial water storage on vertical
deformation rates, but noted the need for independent information from gravity data to resolve an ambiguity between
the two quantities. The periodic deformation transients accompanying the annual hydrological cycle were found by
Bürgmann, Johnson and Dutilleul (annual report) to have similar periodicities to seismicity in the Parkfield area of the
WELCOME
2016 SCEC Annual Meeting page 32
San Andreas fault and the Sierra Nevada/Eastern California Shear Zone. Although the causality and details of the
mechanics of the relationship are still being investigated in detail, these findings motivate further study of hydrologic
transients as a means of understanding the links between surface loading and anthropogenic activity, deformation and
seismicity.
2.3. Other Tectonic Geodesy activities
The estimation of fault slip rates through measurement of interseismic strain accumulation is a mature application of
geodetic data. Deviations from the expected pattern of deformation associated with a uniform, isotropic elastic half
space, as is typically used in such studies, may indicate complexity in crustal elasticity structure. Fialko and colleagues
(annual report) used high resolution GPS and InSAR secular velocity data to investigate the possible presence of a
shear zone with lowered elastic strength in the Anza area of the San Jacinto fault. They found that the anomalously
high shear strain rate present in their data support a 2–3 km-wide shear zone with a strength reduction of 2.4 times,
although the reduction factor trades off against a reduction of the fault locking depth, and the observed strain rate could
also be explained in part by deep fault creep or distributed plastic failure. This finding, if confirmed, could have
consequences for the propagation of seismic waves, and for strong ground motions along the fault zone, and may imply
that the fault segment is reaching the end of its interseismic period.
2.4. Select Publications
● Hammond, B. C., Allen, R. M., Bock, Y., Haase, J. S., Geng, J., Crowell, B. W., & Melgar, D. (2015).
Seismogeodesy of the 2014 Mw6.1 Napa Earthquake, California: Rapid Response and Modeling of Fast
Rupture on a Dipping Strike-slip Fault. Journal of Geophysical Research: Solid Earth, 120(7), 5013-5033. doi:
10.1002/2015JB011921.
● Hollingsworth, J., Leprince, S., Ayoub, F., Dolan, J. F., Milliner, C. W., Tong, X., Sandwell, D. T., & Xu, X.
(2016). Refining the shallow slip deficit. Geophysical Journal International, 204(3), 1843-1862. doi:
10.1093/gji/ggv563.
● Houlié, N., Funning, G. J., & Bürgmann, R. (2016). Use of a GPS-derived troposphere model to improve InSAR
deformation estimates in the San Gabriel Valley, California. IEEE Transactions on Geoscience and Remote
Sensing, 54, 5365–5374.
● Jolivet, R., Simons, M., Agram, P. S., Duputel, Z., & Shen, Z. (2015). Aseismic slip and seismogenic coupling
along the central San Andreas Fault. Geophysical Research Letters, 42, 1-10. doi: 10.1002/2014GL062222.
● Lindsey, E. O., & Fialko, Y. (2016). Geodetic constraints on frictional properties and earthquake hazard in the
Imperial Valley, southern California. Journal of Geophysical Research: Solid Earth,. doi:
10.1002/2015JB012516.
● McGill, S. F., Spinler, J. C., McGill, J. D., Bennett, R. A., Floyd, M., Fryxell, J. E., & Funning, G. J. (2015).
Kinematic modeling of fault slip rates using new geodetic velocities from a transect across the Pacific-North
America plate boundary through the San Bernardino Mountains, California. Journal of Geophysical Research:
Solid Earth, 120(4), 2772-2793. doi: 10.1002/2014JB011459
● Melgar, D., Crowell, B. W., Geng, J., Allen, R. M., Bock, Y., Riquelme, S., Hill, E. M., Protti, M., & Ganas, A.
(2015). Earthquake magnitude calculation without saturation from the scaling of peak ground displacement.
Geophysical Research Letters, 42(13), 5197-5205. doi: 10.1002/2015GL064278.
● Rollins, J. C., Barbot, S. D., & Avouac, J. (2015). Postseismic Deformation Following the 2010 M=7.2 El
Mayor–Cucapah Earthquake: Observations, Kinematic Inversions, and Dynamic Models. Pure and Applied
Geophysics, 172, 1305-1358. doi: 10.1007/s00024-014-1005-6.
● Shen, Z., Wang, M., Zeng, Y., & Wang, F. (2015). Optimal Interpolation of Spatially Discretized Geodetic Data.
Bulletin of the Seismological Society of America, 105(4), 2117-2127. doi: 10.1785/0120140247.
● Spinler, J. C., Bennett, R. A., Walls, C., Lawrence, S., & Gonzàlez-Garcìa, J. (2015). Assessing long-term
postseismic deformation following the M7.2 April 4, 2010, El Mayor-Cucapah earthquake with implications for
lithospheric rheology in the Salton Trough. Journal of Geophysical Research: Solid Earth, 120(5), 3664-3679.
doi: 10.1002/2014JB011613.
● Tong, X., Sandwell, D. T., & Smith-Konter, B. R. (2015). An integral method to estimate the moment
accumulation rate on the Creeping Section of the San Andreas Fault. Geophysical Journal International,
203(1), 48-62. doi: 10.1093/gji/ggv269.
WELCOME
2016 SCEC Annual Meeting page 33
3. Earthquake Geology
The Earthquake Geology Disciplinary Committee promotes studies of the geologic record of the Southern California
natural laboratory that advance SCEC science. Its primary focus is on the Late Quaternary record of faulting and ground
motion, including data gathering in response to major earthquakes. Earthquake Geology also fosters research activities
motivated by outstanding seismic hazard issues, understanding of the structural framework and earthquake history of
special fault study areas, or will contribute significant information to the statewide Unified Structural Representation.
This past year’s accomplishments include:
● Completion of efforts to characterize abrupt uplift and subsidence events associated with major earthquakes
on the Ventura-Pitas Point fault system. These fault ruptures likely spanned a significant length of the western
Transverse Ranges reverse fault system.
● New understanding of multi-fault ruptures stemming from several research fronts, including collaborations
around the Ventura and San Gorgonio Pass Special Fault Study Areas, analysis of the 2010 El Mayor-
Cucapah earthquake sequence, and modeling and paleoseismology of the link between the Imperial fault and
southernmost San Andreas fault.
● Completion and submission for publication of incremental slip rates and paleo-earthquake ages and
displacements for the central Garlock fault, showing abrupt changes in slip rate associated with earthquake
clustering over the past 12,000 years.
● Progress in method development and application of Quaternary geochronometers on fault-related geomorphic
features, as well as progress in assessing their uncertainties and modeling their age distributions.
● Co-sponsoring a workshop on development of a geologic framework for the Community Rheology Mode as a
part of the USR.
Figure 3.1. Data and model of stress state of the El Mayor Cucapah earthquake and its aftershocks in support of the keystone fault hypothesis of Fletcher et al. (2016). a. Regional permuted stress allows the calculation of absolute stress at seismogenic depths. b. Apparent friction of progenitor fault agrees with experimental data, but other faults greatly exceed known strength limits. c. As regional stress builds, slip on optimally oriented faults is regulated by pinning intersections with a misoriented keystone fault. d. Static failure of keystone fault spontaneously spreads to other faults that had reached critical loading earlier in the interseismic cycle.
WELCOME
2016 SCEC Annual Meeting page 34
3.1 Multi-fault ruptures
Geological observations have contributed new understanding of the frequency and mechanisms of multi-fault ruptures.
From analysis of the El Mayor-Cucapah earthquake and its aftershocks, Fletcher et al. (2016) interpret that the network
of faults that ruptured in this event were pinned by a low- to moderate-dip normal fault at depth. The earthquake
nucleated on this ‘keystone fault,’ and propagated upward onto more optimally oriented dextral-normal faults. The
stress state of this earthquake, inferred from a combination of the surface rupture and aftershock data (Figure 3.1),
suggests that most of the faults that slipped in this event were critically stressed at the time of the event. Evidence for
large uplift and subsidence events associated with earthquakes on the Ventura-Pitas point fault support that large,
multi-fault earthquakes may be the dominant mode of strain release along the northern margin of the Ventura-Santa
Barbara basin in the western Transverse Ranges (Rockwell et al., in press; McAulife et al., in review, Simms et al.,
2016 SCEC meeting abstract, Figure 3.2). Conversely, limited fault activity through the San Gorgonio Pass suggests
that most ruptures on the San Andreas fault terminate within the Pass (e.g., Morelan et al., 2016 SCEC abstract). New
efforts to understand coseismic slip variation along the northern Imperial Fault have stimulated a hypothesis that this
fault may link up with large earthquakes on the southernmost San Andreas fault (Meltzner et al. in review, and Oglesby
et al., 2016 SCEC abstract).
WELCOME
2016 SCEC Annual Meeting page 35
Figure 3.2. Evidence for synchronous, coseismic coastal uplift and subsidence of the hanging wall of the Pitas Point-Ventura thrust fault. Top: Lidar topography of Pitas point preserves evidence for three emergent terraces. A fourth, highest terrace is documented from archeological excavations. Figure from Rockwell et al. (in press) Bottom, left: Oxcal model for ages of four marine terrace uplift events since 6.7 ka. The penultimate event, ca 2.1 ka, correlates to an abrupt submergence event at Carpinteria Marsh, age shown in Oxcal model at bottom, right. This could indicate failure of a backthrust and subsidence of its hanging wall during this event (Simms et al., 2016 SCEC annual meeting).
3.2. Understanding links between the San Andreas and ECSZ
One of the most puzzling features of geodetically measured strain in southern California is the transfer of deformation
into the eastern California Shear zone (ECSZ). An excess of 10 mm/yr of dextral shear is measured geodetically across
the little San Bernardino mountains where few active dextral faults are recognized. Two new studies of this area are
helping to understand this problem. Gabriel et al. (2016 SCEC abstract) analyze a new slip-rate site on the Pinto
WELCOME
2016 SCEC Annual Meeting page 36
Mountain fault, a left lateral fault that shunts slip away from the San Gorgonio Pass and into the ECSZ. They use offset
alluvial fans, dated with cosmogenic 10Be, to deduce a slip rate of 2.5 mm/yr over the past 90 kyr. Hislop et al. (2016
SCEC meeting abstract, Figure 3.3) present new geologic mapping of nascent, active dextral faults within the Joshua
Tree region. Slip on one of these faults exceeds 1 km, which is similar to the total slip on several geomorphically well-
expressed faults elsewhere in the ECSZ, and suggests that the paucity of recognizable dextral faulting in the little San
Bernardino mountains may be due to poor preservation of geomorphic evidence and a lack of detailed mapping.
Figure 3.3. Bedrock mapping constraints on young faults associated with historic seismicity in the Black Rock Canyon region of northern Joshua Tree National Park. Inset map shows the location of the study area on a map of southern California Quaternary faults and shaded relief topography. The main map shows Quaternary faults from the USGS fault and fold database with the newly identified Black Rock Canyon fault connecting mapped surface rupture at the south end of the 1992 Landers earthquake sequence (Treiman, 1992) to the Wide Canyon fault. The Black Rock Canyon fault is mapped based upon trends in bedrock faults and fracture patterns, aligned drainages, and subtle scarp-like features. Shaded-relief topography on right is a portion of airborne lidar coverage awarded as an NCALM seed grant to A. Hislop
3.3. Developments in Methodology
Several new developments in methodology have emerged in the past year and throughout SCEC4 in both Quaternary
geochronology and image correlation. On the geochronology front, a new method of dating geomorphic features and
sedimentary deposits based on infrared stimulated luminescence (pIR-IRSL) was developed specifically for potassium
feldspar, which luminesces brighter than quartz and provides better precision than traditional IRSL or OSL (Emmons
et al., 2016 SCEC Abstract). Additionally, efforts to utilize cosmogenic burial dating on offset sedimentary deposits are
ongoing in the San Gorgonio Pass region (e.g. Lifton et al., 2016 SCEC Abstract). Compilations of existing (dominantly
SCEC funded) geochronology from a range of methods (cosmogenic, C-14, U-series, OSL) have yielded new insights
into the uncertainties and best practices associated with dating Quaternary geomorphic features (Gold & Behr, 2016
SCEC Abstract), as well as insights into the geomorphic processes responsible for scatter or discrepancies in age
distributions (Oskin, SCEC Annual Report 2016).
Image correlation algorithms (e.g. COSI-Corr, Milliner et al., 2015 and Aerial2lidar3d, Milliner et al., 2016 SCEC
Abstract) from pre- and post-event aerial photos and LiDAR scans have produced significant advancements in
quantifying coseismic surface deformation for recent earthquakes within the SCEC domain. Results from the eastern
WELCOME
2016 SCEC Annual Meeting page 37
California shear zone reveal new insight into the balance of fault slip and distributed deformation of the surrounding
rock volume. Reanalysis of coseismic slip from the 1992 Landers earthquake (Milliner et al., 2015, Milliner et al., 2016)
and the 1999 Hector Mine earthquake (Sousa et al., in review) confirm sharp, along-strike slip gradients near 10^-3
(1m in 0.5 km), and suggest that fault slip distributions could be fractal rather than smoothly elliptical. Work is also
ongoing to characterize coseismic surface deformation patterns associated with the 2010 El Mayor-Cucapah EQ (Lajoie
& Nissen, 2016 SCEC Abstract) and Mw 7.3 1952 Kern County EQ along the White Wolf fault (Hatem et al, 2016 SCEC
Abstract).
3.4. Select Publications
● Dolan, J. F., McAuliffe, L. J., Rhodes, E. J., McGill, S. F., & Zinke, R. (2016). Extreme multi-millennial slip rate
variations on the Garlock fault, California: Strain super-cycles, potentially time-variable fault strength, and
implications for system-level earthquake occurrence. Earth and Planetary Science Letters, (in press). SCEC
Contribution 6243.
● Fletcher, J. M., Oskin, M. E., & Teran, O. J. (2015). The role of a keystone fault in triggering the complex El
Mayor-Cucapah earthquake rupture. Nature Geoscience, 9, p. 303–307 doi:10.1038/ngeo2660. SCEC
Contribution 1962.
● McAuliffe, L. J., Dolan, J. F., Rhodes, E. J., Hubbard, J. A., Shaw, J. H., & Pratt, T. L.(2015). Paleoseismologic
evidence for large-magnitude (Mw≥7.5) earthquakes on the Ventura blind thrust fault: Implications for multi-
fault ruptures in the Transverse Ranges of southern California. Geosphere, (under review). SCEC Contribution
2076.
● Meltzner, A. J., Rockwell, T. K., & Tsang, R. (2015). A long record of surface ruptures on the northern Imperial
fault, including slip in the past six events. Bulletin of the Seismological Society of America, (under review).
SCEC Contribution 6060.
● Milliner, C. W. D., J. F. Dolan, J. Hollingsworth, S. Leprince, F. Ayoub, and C. G. Sammis (2015), Quantifying
near-field and off-fault deformation patterns of the 1992 Mw 7.3 Landers earthquake, Geochem. Geophys.
Geosyst., 16, 1577–1598, doi:10.1002/2014GC005693.
● Milliner, C.W.D., Sammis, C., Allam, A.A., Dolan, J.F., Hollingsworth, J., Leprince, S., Ayoub, F., (2016).
Resolving Fine-Scale Heterogeneity of Co-seismic Slip and the Relation to Fault Structure. Scientific Reports
6, 27201.
● Sousa, F. J., Stock, J. M., Scharer, K. M., Hudnut, K. W., Akciz, S. O., Witkosky, R. D., & Harvey, J. C. (2015).
Re-evaluating offset measurements in the maximum slip zone of the 1999 Hector Mine Earthquake surface
rupture. Bulletin of the Seismological Society of America, (under review). SCEC Contribution 6121.
WELCOME
2016 SCEC Annual Meeting page 38
4. Computational Science (CS)
The Computational Science Disciplinary Group promotes the use of advanced numerical modeling techniques and high
performance computing (HPC) to address the emerging needs of SCEC users and the application community on HPC
platforms.
This past year’s accomplishments include:
● The fast multipole method was used to speed up boundary element calculations. Features of the new
boundary element code include topography, material heterogeneity, plasticity, and gravity. In addition, neural
networks were used to obtain more than three orders of magnitude speed-up in viscoelastic models.
● A similarity search algorithm for detect repeating earthquakes in continuous waveform data, termed
Fingerprint and Similarity Thresholding (FAST), provides orders of magnitude speed-up relative to techniques
like autocorrelation.
● The SCEC wave propagation and rupture dynamics code AWP-ODC was released under BSD-2 clause open-
source license. A version of this code accounting for nonlinear material response was used for 0-4 Hz Mw7.7
San Andreas Fault scenario earthquake wave propagation simulation.
● Significant progress was made this year, with Intel Parallel Computing Center support, to adopt SCEC HPC
codes, including AWP-ODC and discontinuous Galerkin-based EDGE, to next generation Xeon Phi KNL
based architectures.
● New finite difference and finite element codes were developed for modeling earthquake sequences with rate-
and-state friction, accounting for viscoelastic, poroelastic, and elastic-plastic off-fault material response and
fault roughness. Similarly, PyLith, an open-source community code for quasi-static and dynamic faulting
problems, has been extended using a numerical technique that permits pointwise implementation of a wide
range of physics (e.g., fault friction laws, bulk rheologies, and even additional governing equations).
4.1 Data-intensive HPC revolutionizes earthquake detection in continuous waveform data
Massive amounts of continuous seismic data are routinely
collected for earthquake monitoring purposes. Classic event
detection methods like LTA/STA fail to detect small events that
are not well above the noise level. Approaches like template
matching and autocorrelation either require prior knowledge of
events (i.e., templates) or are inefficient computationally.
Drawing on techniques from data-intensive computing, Yoon
et al. (2015) have devised a new method, Fingerprint and
Similarity Thresholding (FAST), with high sensitivity,
applicability, and computational efficiency (Figure 4.1). FAST
exploits a sparse representation of the data, or fingerprints,
then uses locality sensitive hashing to search efficientlyfor
fingerprint matches.
HPC has also been used to enable event detection using
template matching on XSEDE GPU clusters. Meng and Peng
(2016) applied this approach to waveform data in Southern
California, focusing on the aftershock sequence following a
Mw 5.1 event on the San Jacinto fault. The magnitude of
completeness was reduced from 1.2 to 0 using this new
technique.
4.2. Open-source release of SCEC wave propagation code AWP-ODC
The SCEC AWP-ODC code for seismic wave propagation and rupture dynamics has been extensively utilized for many
of SCEC’s largest supercomputer simulations. This year the code, AWP-ODC-OS, was released for the first time under
an open-source licence (BSD-2 clause license) and continuous updates are provided to users (Mu et al., 2016). AWP-
ODC-GPU has been further optimized for scalability and efficiency. By utilizing the multi-streaming technique, we have
been able to effectively overlap communication with computation, with 100% of parallel efficiency achieved on 16,384
XK7 GPUs with the linear code. To our knowledge, this is a new scaling record on GPU-based architectures. The
WELCOME
2016 SCEC Annual Meeting page 39
computational cost of the nonlinear code, which now integrates frequency-dependent attenuation and plasticity, was
reduced from initially 4x to less than 50% with respect to the highly efficient linear code, with 2.9 Pflop/s recorded in
benchmarks on OLCF Titan (Roten et al., 2016). In addition, a scalar version of discontinuous mesh based AWP-ODC
has been parallelized and currently being tested for stability through a collaboration between SDSC and SDSU. The
primary goal of this project is to include the near-surface low-velocity material in simulations of large earthquakes up to
higher frequencies than previously possible, and improve parallel performance by about an order of magnitude.
In addition, a hybrid version of AWP-ODC is being developed for Xeon Phi processors. Initial performance results
demonstrate that this extension running on an Intel Xeon Phi 7210 processor performs more than 7.5 times faster than
on an Intel Xeon E5-2680v3 and more than 1.4 times faster than on the recent Nvidia M40 GPU. Furthermore, we have
developed a novel local time stepping (LTS) scheme for the ADER-DG finite element method and a holistic optimization
targeting seismic simulations on the Intel Xeon Phi x200 processor. Our LTS-scheme summarizes elements with similar
time steps and thus aims for optimal large-scale throughput by trading some of the algorithmic optimality. The
performance comparisons demonstrate that KNL is able to outperform its previous generation, the Intel Xeon Phi
coprocessor x100 family, by more than 2.6× in time-to-solution when running LTS-workloads. Additionally, our results
show a 2.9× speedup compared to latest Intel Xeon E5v3 CPUs. Our performance benchmarks (4th order, matrix Fm2,
M=9, N=20, K=20, 16% non-zero, hot L2) underline the potential of the approach and show that we are able to reduce
time-to-solution by more than 3.5x on the Intel Xeon Phi processor as compared to the fastest non-fused
implementation (Breuer et al., 2016). SCEC KNL development is supported by Intel Parallel Computing Center program
4.3. Improvements to Earthquake Simulators
The earthquake simulator code RSQSim has been ported to the NSF supercomputing systems including NCSA Blue
Waters and TACC Stampede. Significant optimization has been achieved at lower end of the core counts. Run-time
adjustments were implemented for effective use of the systems, and MPI-IO was added for both initialization and output
writing. Work is underway on algorithmic changes that would streamline the calculation and reduce or remove the
computational load imbalance.
Additional work was completed to enable integration and use of the SCEC Community Fault Model (CFM) into
earthquake simulator codes. First, a version of the CFM was created with more equidimensional triangular fault
elements (essential for numerical accuracy). The upper edge of surface-breaking faults, which encounter nonplanar
free-surface topography, was flattened for compatibility with earthquake simulators (which presently assume a flat free
surface). Finally, slip rate and rake directions, which must be included for earthquake simulators but are not provided
with the CFM, were obtained from additional sources. This project is nearing completion and will permit tight integration
of the CFM into earthquake simulators.
4.4. Select Publications
● Roten, D., Y. Cui, K.B. Olsen, S. M. Day, K. Withers, W. Savran and P. Wang (2016). "High-frequency
Nonlinear Earthquake Simulations on Petascale Heterogeneous Supercomputers", SC16, November 13-18,
2016, Salt Lake City, UT (in press).
● Roten, D., K.B. Olsen, S. M. Day and Y. Cui (2016). Quantification of fault zone plasticity effects with
spontaneous rupture Simulations, PAGEOPH, submitted.
● Roten, D., K. B. Olsen, Y. Cui, and S. M. Day (2015), Quantification of fault zone plasticity effects with
spontaneous rupture simulations, Workshop on Best Practice in Physics-Based Fault Rupture Models for
Seismic Hazard Assessment of Nuclear Installations, Vienna, Austria, Nov 18-20.
● Heinecke, A., A. Breuer, M. Bader, High Order Seismic Simulations on the Intel Xeon Phi Processor (Knights
Landing), International Supercomputing Conference (ISC) 2016.
● Breuer, A., A. Heinecke, M. Bader, Petascale Local Time Stepping for the ADER- DG Finite Element Method,
IPDPS 2016.
● Heinecke, A., A. Breuer, M. Bader, Chapter 21 - High Performance Earthquake Simulations, Intel Xeon Phi
Processor High Performance Programming Knights Landing Edition, Elsevier, 471-483, 2016.
● Meng, X., and Zhigang Peng (2016) Increasing lengths of aftershock zones with depths of moderate-size
earthquakes on the San Jacinto Fault suggests triggering of deep creep in the middle crust, Geophys. J. Int.,
204 (1): 250-261, doi:10.1093/gji/ggv445.
● Yoon, C. E., O. O’Reilly, K. J. Bergen, and G. C. Beroza (2015) Earthquake detection through computationally
efficient similarity search, Science Advances, 1(11), e1501057, doi: 10.1126/sciadv.1501057.
WELCOME
2016 SCEC Annual Meeting page 40
5. Unified Structural Representation (USR)
The Unified Structural Representation (USR) focus group develops models of crust and upper mantle structure in
California for use in a wide range of SCEC science, including ground-motion prediction, earthquake hazards
assessment, and fault systems analysis. These efforts include the development of Community Velocity Models (CVM's)
and Community Fault Models (CFM's), which when integrated together comprise Unified Structural Representations
(USR’s). In partnership with other working groups in SCEC, the USR Focus Area also helps support the evaluation and
improvement of these models through ground-motion simulations, 3D waveform tomography, earthquake relocations,
and fault systems modeling.
Given the anticipated transition to SCEC5, this past year’s principal fault modeling efforts focused on completing major
upgrades to the community models and enhancing the systems that make these resources accessible:
● SCEC Community Fault Model (CFM): Development and release of a new version of the CFM (v. 5.1) for
southern California that fully incorporates the refinements to fault representations based on relocated
seismicity catalogs and detailed fault traces in the USGS Quaternary Fault and Fold database. This version
represents completion of a multi-year effort to incorporate major enhancements of the fault representations in
the San Andreas system, the Peninsular Ranges, the Mohave region, and Western Transverse Ranges. These
refinements include the faults systems that are the focus of the SCEC Special Fault Study Areas. The model
also fully implements a new hierarchical
naming and numbering system that is
compatible with the USGS Fault and Fold
Database.
● Central California USR (CFM):
Development of the first complete Unified
Structural Representation for Central
California (USR-CC) that incorporates major
basin structures and fault representations.
This new model supports SCEC's Central
California Seismic Project (CCSP), which is
using this structural representation to
facilitate 3D waveform tomography studies.
In addition, the USR Focus Area continues to support
existing Community Models, including the Statewide
Community Fault Model (SCFM) (Figure 5.1) and
southern California Unified Structural Representation
(USR).
5.1 Community Fault Model (CFM)
SCEC has completed a major effort to systematically refine the Community Fault Model (CFM) in southern California
(Plesch et al., 2007) using detailed fault traces from the USGS Quaternary Fault & Fold Database, precisely relocated
earthquake hypocenters, and new focal mechanism catalogs (e.g., Yang et al., 2012; Hauksson et al., 2012). This
results in fault representations that are more precise, and often more highly segmented than in previous model versions.
These refinements were completed over the course of three years, with an initial focus on the San Andreas system
(e.g., Nicholson et al., 2016; Fuis et al., 2016), Peninsular Ranges, and Mohave region. This past year, we completed
a series of refinements to the fault representations in the Western Transverse Ranges and Santa Barbara Channel
(Figure 5.2). This region contains a complex system of imbricated thrust and reverse faults, several of which are blind,
and encompasses the Ventura Special Fault Study Area. More precise fault representations were developed by
mapping in dense grids of 2- and 3-D marine seismic reflection data and by incorporating relocated earthquake
hypocenters and focal mechanisms from several recent earthquake sequences (Sorlien et al., 2016; Nicholson et al.,
2016). Given the diffuse nature of much of the seismicity in this region, however, the mid-crustal geometry for many of
these major fault systems remains unconstrained. Thus, the latest model release (CFM 5.1, Plesch et al., 2016)
includes alternative representations of several major faults in this region. For example, the Ventura – Pitas Point fault
system is represented both with and without a mid-crustal detachment level (Hubbard et al., 2014; Sorlien et al., 2016;
WELCOME
2016 SCEC Annual Meeting page 41
Nicholson et al., 2016). CFM 5.1 representations are available in their native tsurf formats, as well as in rectilinear and more uniformly discretized tsurf versions.
To improve the systems that make these CFM models available, we initiated a project in collaboration with the SCEC Community Modeling Environment (CME) team to develop a new relational database structure and web interface that will allow users to identify, assemble, and download model components. When completed, these new tools will enable users to search the database by fault name, geographic region, attributes (e.g., fault type), and other criteria. To facilitate this future database system, the latest version of the CFM implements a hierarchical naming system that includes fault zones, sections, names, and unique identification numbers. This classification system is compatible with the U.S.G.S. Fault and Fold Database.
5.2. Central California USR This past year we developed and refined a Unified Structural Representation (USR) (Shaw et al., 2015) for central California in support of the Central California Seismic Project (CCSP). This model includes new representations of major basin structures and faults from the southern California and Statewide Community Fault Models (CFM and SCFM). The model is intended to help facilitate a more comprehensive assessment of earthquake sources and physics-based waveform modeling to define path effects in the region.
This effort involved the development of a new velocity model of the San Joaquin basin using a large database of well logs, seismic reflection profiles, and geological data that constrain basin Vp, Vs, and density structure. The new model includes a detailed description of the basin shape, represented by top-basement, base Tertiary, base Quaternary, and other geologic horizons. These surfaces are constrained by petroleum wells and seismic reflection profiles, as well as geologic cross sections and maps from previous studies in the region. These surfaces are compatible with the positions and offsets of major faults in the Statewide Community Fault Model (SCFM) (Figure 5.1). Incorporating these fault structures helps to define aspects of basin shape that may lead to seismic wave focusing and amplification.
WELCOME
2016 SCEC Annual Meeting page 42
The velocity and density structure within the basin was constrained by a rich database of sonic logs and stacking
velocities from industry reflection profiles. Sonic logs and stacking velocities measure compressional wave speeds
(Vp), and were used to parameterize the sediment velocities. A Vp/Vs relationship was calibrated using dipole sonic
log data, allowing us to generate a co-registered Vs model. The resulting velocity parameterizations were implemented
within the enveloping topographic/ bathymetric and top-basement surfaces to develop a basin wavespeed model by
the procedure illustrated in Figure 5.3. This new model, along with refined versions of other basin models (e.g., Santa
Maria basin), were embedded in a regional velocity model (6th iterate of CCA; Chen et al., 2016) to develop a
preliminary description of crustal velocity structure in the CCSP study area. Subsequent tomographic iterations of this
model will further improve our knowledge of 3D wavespeed structure, thereby helping to facilitate more comprehensive
assessment of earthquake sources and physics-based waveform modeling to define path effects in the region.
5.4 Community Seismic Velocity Model
Several research efforts targeted specific regions for improvement in the characterization of the elastic properties.
Additional work will be necessary to incorporate these advances into the community seismic velocity models.
Asimaki, Shi, and Taborda developed a model for generating shallow shear wave velocity versus depth profiles from
Vs30 for use in wave propagation-based ground-motion simulations. The model also includes a spatially correlated
stochastic component. They validated the model using observations of site amplification. This model can be used to
populate seismic velocity models at shallow depths by bridging the gap between Vs30 measurements and coarser
resolution constraints, such as those from full-waveform inversions.
Thurber and others continued their efforts to improve the characterization of the elastic properties in the San Jacinto
Fault Zone region. Their current work increases the resolution via joint inversions of body-wave and surface-wave
datasets rather than separate inversions. Based on a time-frequency goodness-of-fit criterion, they found that the model
reproduces body wave arrivals at frequencies up to 3 Hz but does not do as well for high-amplitude surface waves,
especially at frequencies above 1 Hz. This model shows some improvement over CVM-H in modeling regional
waveforms.
Share and others focused on constraining the seismic velocity structure for the San Gorgonio Pass Special Fault Study
Area. They used fault-zone headwaves and double-difference travel-time tomography to identify spatial variations in
elastic properties, including two material contrasts across the San Andreas Fault. One material contrast extends from
slightly northwest of Cajon Pass to the vicinity of San Gorgonio Pass with a slower seismic velocity on the southwest
side of the San Andreas Fault; this favors ruptures propagating to the northwest. The second velocity contrast extends
from slightly northeast of San Gorgonio Pass to the Coachella Valley with the opposite velocity contrast; this favors
ruptures propagating to the southeast.
5.5. Select Publications
● Fuis, G. S., Catchings, R. D., Scheirer, D. S., Goldman, M. R., & Zhang, E. (2016). Structure of the San
Andreas Fault Zone in the Salton Trough Region of Southern California: A Comparison with San Andreas
Fault Structure in the Loma Prieta Area of Central California. Poster Presentation at 2016 SCEC Annual
Meeting.
● Nicholson, C., Sorlien, C. C., & Hopps, T. E. (2016). Anomalous Uplift at Pitas Point: Implications from
Onshore & Offshore 3D Fault & Fold Geometry and Observed Fault Slip. Poster Presentation at 2016 SCEC
Annual Meeting.
● Plesch, A., Nicholson, C., Sorlien, C. C., Shaw, J. H., & Hauksson, E. (2016). CFM Version 5.1: New and
revised 3D fault representations and an improved database. Poster Presentation at 2016 SCEC Annual
Meeting.
● Shaw, J. H., A. Plesch, C. Tape, M. P. Suess, T. H. Jordan, G. Ely, E. Hauksson, J. Tromp, T. Tanimoto, R.
Graves, K. Olsen, C. Nicholson, P. J. Maechling, C. Rivero, P. Lovely, C. M. Brankman, J. Munster, 2015,
Unified Structural Representation of the southern California crust and upper mantle, Earth and Planetary
Science Letters, 415, 1–15.
● Sorlien, C. C., Nicholson, C., Behl, R. J., & Kamerling, M. J. (2016). Displacement direction and 3D geometry
for the south-directed North Channel – Pitas Point fault system and north-directed ramps, decollements, and
other faults beneath Santa Barbara Channel. Poster Presentation at 2016 SCEC Annual Meeting.
WELCOME
2016 SCEC Annual Meeting page 43
6. Fault and Rupture Mechanics (FARM)
The Fault and Rupture Mechanics (FARM) interdisciplinary group focuses on understanding earthquake rupture
mechanics through a combination of theoretical modeling, laboratory experiments and field observations. The results
from FARM research are closely linked to efforts in the SDOT, UCERF, and CSM programs (among others) in SCEC4.
Improvements in computational capabilities are making it possible to more properly model dynamic rupture propagation
on geometrically realistic fault structures. Similarly, technical advances in experimental and analytical equipment are
opening new opportunities for investigating earthquake deformation processes during quasi-static and dynamic
conditions in both laboratory and natural fault samples. Progress in this area remains diverse and projects are
numerous; however, several themes remain at the forefront as we look forward to SCEC5. Details on these projects
(referenced by PI name without year) are found in the project reports from 2015.
6.1 Heterogeneous Fault Stress and
Structure
Considerable effort remained focused on how
heterogeneous fault stress and fault structure (e.g.,
roughness, larger scale fault segmentation and
geology) influence seismicity and rupture propagation.
Several studies have explored the role of fault
roughness on earthquake processes, and the geologic
record of these structures.
Cooke et al. conducted quasi-static crustal deformation
models that do not include material heterogeneities
(Figure 6.1). In regions of fault complexity, residual off-
fault stresses produce permanent deformation
including microseismicity. The spatial correlation of
patterns of stresses from focal mechanism inversions
with off-fault stresses over multiple earthquake cycles
suggests that the microseismicity in the San Gorgonio
Pass region records permanent off-fault deformation
that is different from the long-term strike-slip loading of
the San Andreas. Thus, stress inversions from
microseismicity catalogs may provide inaccurate
information about the loading of the major faults.
Dynamic rupture simulations on heterogeneous, rough
faults have been developed to quantify the frictional
weakening length scale using observed off-fault peak
velocity as a proxy (Daub); the estimated slip weakening distance is related to fault roughness, with the highest values
occurring near restraining bends. Biasi developed the first empirical relation for the relative ability of fault bends to stop
rupture as a function of bend angle. The observed relationship has applications in PSHA and informs dynamic and
kinematic rupture models. Olsen developed a rupture generator that includes the higher frequency energy from
geometrical roughness, frequency-dependent attenuation, and small-scale heterogeneities. These dynamic rupture
models, for rough faults with elasto-plastic off-fault yielding, show that regions with higher stress-drop correspond to
larger variations in slip, slip rate, and rupture speed.
To understand complexity in source process and its expression in high frequency ground motion, Dunham completed
simulations of dynamic ruptures on fractally rough faults, with strongly rate-weakening friction laws and off-fault
plasticity, focusing on supershear transition in 2D models and 3D simulations. Supershear transitions initiate at
geometric complexities and become more likely on rougher faults. However, sustained supershear propagation is
favored on locally smooth fault segments. To design laboratory experiments that reproduce numerical results of
supershear rupture triggering and transition on low-pre-stress heterogeneous faults, Lapusta developed a method of
imaging local friction on experimental faults during dynamic rupture propagation. The results demonstrate that linear
slip-weakening friction is an inadequate description; the friction response can be qualitatively explained by the
WELCOME
2016 SCEC Annual Meeting page 44
combination of rate-and-state friction with flash-heating-like co-seismic weakening (Figure 6.2). Viesca and colleagues
examined the extent that spatial variations in frictional properties determine how and where an earthquake-generating
dynamic fault rupture may initiate (Viesca 2015; 2016).
Rockwell et al. analyzed paleoseismic observations that indicate that the southern San Andreas and Imperial faults
(SAF and IF) ruptured at similar times in past earthquakes, suggesting that they ruptured together during single events.
Modeling of dynamic rupture scenarios suggests that large earthquakes may rupture through this large step-over,
depending on rupture directivity and the presence of connecting faults or cross-faults, supporting the paleoseismic
observations. Nucleation on the SAF favors rupture of both the SAF and IF, whereas nucleation on the IF ruptures
primarily the western (IF) strand of the step-over.
6.2 Off-fault Plasticity
Ampuero and Mao (2016) developed a fracture mechanics model and large-scale dynamic rupture simulations with off-
fault plasticity to explain the initial growth and eventual saturation of damage zone thickness as a function of fault
displacement. They found that fault zone thickness is limited by the finite seismogenic depth, which limits the stress
concentration at rupture tips. In contrast, the absence of saturation of the thickness of splay fault fans was attributed to
the effects of stress relaxation in the underlying asthenosphere, which is insensitive to seismogenic depth. The depth
extent of such fault damage zones is often poorly resolved by geophysical data. Li (2016) shows that analysis of fault
zone trapped waves generated by teleseismic waves can constrain the deep structure of active fault zones, providing
insights on the depth of damage along the Calico fault. Erickson developed an efficient, computational framework for
simulating multiple earthquake cycles with off-fault plasticity. Off-fault plastic flow ensues when stresses reach the yield
surface, generating a damage zone that evolves with each event (in which up to ~1 m of offset is accommodated by
inelastic deformation after 10 ruptures).
Dolan developed a methodology to investigate surface deformation patterns of large earthquakes utilizing differencing
of post-event lidar data with pre-event digital elevation models (DEMs). Results for the Hector Mine earthquake show
vertical motion ranging from 0.40 to 1.35 m, in agreement with field data. However, they found significant differences
in places of fault structural complexity, indicating the presence of off-fault deformation.
6.3 Dynamic Weakening and Frictional Properties
Understanding the frictional processes responsible for dynamic weakening mechanisms remains a focus of
experimental and geological studies. Laboratory experiments characterize the processes responsible for thermal
pressurization. Tullis et al. developed protocols to isolate and characterize thermal pressurization in controlled
laboratory experiments. Theoretical studies provide new insights into the physical processes responsible for dynamic
weakening, and rationale for their inclusion into earthquake cycle and rupture models. Carlson applies STZ theory to
experiments of granular shear at very low normal stress with the aim of ultimately extrapolating results to the conditions
of natural earthquakes. They demonstrate successful models of high speed friction experiments with vibration (auto-
acoustic compaction) and conducted simulations at higher stress to map out transitions among stick-slip, vibration,
WELCOME
2016 SCEC Annual Meeting page 45
confining stress, and sliding rate. Elbanna et al. apply STZ theory in 2D simulations that allows grain fracture, for
studying localized shear at elevated stress and the onset of brittle behavior. The "brittle-ductile" transition in these
layers is controlled by porosity. Evans and colleagues developed a new fault surface paleothermometer that exploits
the temperature-sensitive oxidation state changes in transition metals that result from shear heating during seismic slip.
In addition, they developed methods to date the slip on faults using UTh/He geochronology and used XANES analyses
of natural fault surfaces from the San Andreas system to estimate frictional ΔT of ~450°C (Figure 6.3).
Lapusta explores the possibility of large earthquake slip extending into the deeper creeping fault extensions and its
effect on microseismicity and fault coupling by numerical simulations of long-term slip in several models, including a
model with dynamic weakening (DW) restricted to the velocity-weakening (VW) region and a model with DW extending
deeper into velocity-strengthening (VS) regions. The second model leads to deeper rupture. These results also illustrate
how microseismicity at the base of the seismogenic zone is suppressed in models with deeper slip penetration.
6.4 Fault Properties
Based on mapped fault structures and seismicity data, Brodsky created flow models to constrain the extent and
amplitudes of injection-induced pore-pressure perturbations. Using a range of realistic damage zone widths and
permeability values, they showed that induced pore-pressure changes can produce seismic activity up to ~10 km away
from injection wells (Goebel et al. 2016). Allegre et al., (2015) made new in situ measurements of scale-dependent
hydrogeological properties within fault zones utilizing a combination of tidal, barometric and seismic response analyses
on pressure head time series from the Santa Susana Field Laboratory. Permeability computed from tidal response is
homogeneous over the site, regardless of the presence of faults. The tidal responses showed higher variability of
specific storage inside the fault zones, suggesting that the fault damage generates a fluid storage architecture.
Barbot explores deterministic models to reproduce period doubling sequence observed in a single low frequency
earthquake family. The sequence is disrupted by the 2004 Parkfield earthquake; the model includes a single asperity
with homogenous frictional properties and a pore pressure change is imposed at the time of the Parkfield earthquake.
Ryan et al (2015) developed dynamic rupture and tsunami simulations of M7.7 earthquakes on the Pitas Point and
Lower Red Mountain faults in the Ventura basin. The modeled inundation in Ventura exceeds the estimates currently
used by emergency managers. The inundation map is not significantly sensitive to whether the ruptures are blind or
reach the surface.
6.4. Select Publications
● Allegre, V., E.E. Brodsky, B. L. Parker, and J. A. Cherry (2015), Field Evidence for a Low Permeability, High
Storage Fault Core at the Santa Susana Field Laboratory, Fall 2015 AGU, MR33A-2627.
● Ampuero, J.P., and X. Mao (2016), Upper limit on damage zone thickness controlled by seismogenic depth.
To appear in AGU Monograph "Fault Zone Dynamic Processes: Evolution of Fault Properties During Seismic
Rupture".
● T.H.W. Goebel, S.M. Hosseini, F. Cappa, E. Hauksso, J.-P. Ampuero, F. Aminzadeh & J.B. Saleeby (2016).
“Wastewater disposal and earthquake swarm activity at the southern end of the Central Valley, California”,
Geophys. Res. Letts., 43, 1-8, doi:10.1002/2015GL066948.
WELCOME
2016 SCEC Annual Meeting page 46
● Li, Y.G. (2016), Using fault-zone trapped waves from teleseismic earthquakes to view deep structure of the
Calico fault in Southern California, submitted as a book chapter.
● Ryan, K. J., E. L. Geist, M. Barall, and D. D. Oglesby (2015), Dynamic models of an earthquake and tsunami
offshore Ventura, California, Geophysical Research Letters, 42.
● Viesca, R. C. (2016), Stable and unstable development of an interfacial sliding instability, Phys. Rev. E., 93(6),
060202(R).
● Viesca, R. C. (2016), Self-similar slip instability on interfaces with rate- and state- dependent friction, Proc.
Roy. Soc. A, 472(2192), 20160254.
WELCOME
2016 SCEC Annual Meeting page 47
7. Southern San Andreas Fault Evaluation (SoSAFE)
The SCEC Southern San Andreas Fault Evaluation (SoSAFE) group has aimed to increase knowledge of slip rates,
paleoearthquake ages, and slip distributions of past earthquakes, for the past two thousand years on the southern San
Andreas fault (SAF) system. From Parkfield to Bombay Beach, and including the San Jacinto fault, the objective is to
obtain new data to clarify and refine relative hazard assessments for each potential source of a future 'Big One'.
This past year’s accomplishments include:
● A combined paleoseismic record for the last 1100 years for the Salton Sea region, covering the southern-most
SAF, southern San Jacinto, and Elsinore faults that shows variations in moment release rate (Figure 7.1).
This will be highlighted at the SCEC Annual Meeting in a plenary presentation by Tom Rockwell.
● New and promising paleoseismic
site on the less-studied Cholame
section of the SAF that will help to
connect paleorupture models from
the Carrizo through Parkfield to the
creeping section (Figure 2).
● Pleistocene and Holocene slip
rates along the Mission Creek
strand of the SAF and Holocene
slip rates along the Banning strand
and San Gorgonio Pass thrust fault
are internally consistent and show
moderate slip rates (<10 mm/yr) on
the southern side of the pass.
● Dynamic rupture models explore
the effect of fault geometry, stress
conditions and hypocenter location
on through-going ruptures. Paired
with paleoseismic data, these new
results suggest that some longer, multisegment ruptures occurred in the past and that these ruptures should
be considered in hazard planning efforts.
7.1 Paleoearthquake Ages and
Displacements
Exciting new work is underway on the
southern SAF that provides data on the size
and timing of prehistoric large earthquakes.
Rockwell (2016 SCEC plenary talk)
completed an effort to determine a Salton
Trough-wide earthquake history utilizing the
depositional records of Lake Cahuilla (Figure
7.1). The advantage of using the Lake
Cahuilla record is that a clearly identified set
of regionally distributed isochronous markers
associated with rising and falling lake levels is
disrupted by surface ruptures. The regional
record suggest that earthquake timing and
moment rate has been variable in the past
1100 years, and on the southern SAF, San
Jacinto, and Elsinore faults moment release rate has been lower than average for the two centuries. An intriguing
possibility suggested from the research is that earthquake recurrence is modulated by lake loading—suggesting a fine
sensitivity to those forces.
WELCOME
2016 SCEC Annual Meeting page 48
Fieldwork and focused geochronology in 2016 is proving new data on the earthquake history of the northern end of the
southern SAF. On the Carrizo section, at Van Matre Ranch, preservation of gullies offset by the fault in the last 4
earthquakes indicates that slip may be variable along strike, with 4 m of slip in 1857 and either 8 m in the penultimate
event or 8 m in the last three prehistoric events (Salisbury et al., 2016 SCEC meeting abstract). On the Cholame
section, the new LT paleoseismic site shows evidence of several paleoearthquakes in a section with abundant charcoal
that will facilitate dating (Williams et al. 2016 SCEC meeting abstract) (Figure 7.2).
7.2. San Andreas Fault Slip Rate Studies
A major focus of SoSAFE and related San
Gorgonio Pass Special Fault Study Area
fieldwork is the examination of variation in slip
rates along major strands of the southern SAF.
From southeast to northwest, these include
studies showing rapid slip on the Mission Creek
strand over both the late Pleistocene (~20-25
mm/yr, near Highway 62, Blisniuk et al. SCEC
Final Report #15127 and at Pushawalla), and
in the Holocene (16-22 mm/yr, at the Three
Palms site, Munoz et al., 2016 SCEC meeting
abstract and at Pushawalla, Blisniuk et al. 2015
SCEC Annual Meeting). These observations
are consistent with relatively slow (<7 mm/yr)
Holocene slip rates obtained on both ends of
the Banning strand (Gold et al., 2015; Scharer
et al., AGU 2015). No slip rate sites have been
found along the Garnet Hill strand, but geologic
mapping by Heurta et al. (2015 SCEC meeting)
show that slip on the Banning is transferred
onto the San Gorgonio Pass thrust fault near Whitewater. Thus the low rates on the Banning strand and the San
Gorgonio Pass thrust fault (Heermance et al., 2015 SCEC meeting abstract) suggest the Garnet Hill strand does not
carry significant strain.
7.3. Pairing Dynamic Rupture Models and Paleoearthquake Rupture Scenarios
A new method for testing the permissibility of paleoearthquake rupture models (e.g. Scharer et al., 2016; Bemis et al.,
2016) is to use dynamic rupture modeling to explore how fault geometries, stress conditions, and hypocentral locations
affect rupture length and ultimately, earthquake magnitude. A study of the geometry of the San Jacinto and San
Andreas Faults, for example, suggests that the historic 1812 earthquake may have ruptured from the San Bernardino
strand of the SAF onto the northern portion of the San Jacinto Fault (Lozos, 2016; Figure 7.3). Similar dynamic rupture
modeling of the San Andreas and Imperial Fault system indicates that south-directed ruptures are favored under a
wider range of parameters and permits that some of the paleoseismic events could be the same rupture (Oglesby et
al., and Meltzner et al., 2016 SCEC meeting abstract). These types of studies provide independent tests of the UCERF3
rupture catalogue, and indicate that longer ruptures that will affect a larger region of southern California should be
considered for hazard planning and mitigation efforts.
7.4. Select Publications
● Bemis, S., Scharer, K., Dolan, J., Rhodes, E., 2016. The Elizabeth Lake paleoseismic site: Rupture pattern
constraint for the past ~800 years for the Mojave section of the south-central San Andreas Fault, 7th
International INQUA Meeting on Paleoseismology, Active Tectonics, and Archeoseismology, 277-280.
● Lozos, J. (2016). A case for historic joint rupture of the San Andreas and San Jacinto Faults, Science,
http://dx.doi.org/10.1126/sciadv.1500621
● Rockwell, T. K. (2016). Open Intervals, Clusters and Supercycles: 1100 years of Moment Release in the
Southern San Andreas Fault System: Are we Ready for the Century of Earthquakes?. Oral Presentation at
2016 SCEC Annual Meeting.
WELCOME
2016 SCEC Annual Meeting page 49
● Scharer, K., Weldon, R., and Bemis, S., 2016. Testing geomorphology derived rupture histories against the
paleoseismic record of the southern San Andreas fault, 7th International INQUA Meeting on Paleoseismology,
Active Tectonics, and Archeoseismology, 216-219.
WELCOME
2016 SCEC Annual Meeting page 50
8. Stress and Deformation Over Time (SDOT)
The focus of the interdisciplinary focus group Stress and Deformation Over Time (SDOT) is to improve our
understanding of how faults are loaded in the context of the wider lithospheric system. SDOT studies these processes
on timescales from tens of millions of years to tens of years using the structure, geological history, and physical state
of the southern California lithosphere as a natural laboratory. The objective is to tie the present-day state of stress and
deformation on crustal-scale faults and the lithosphere as a whole to the long-term, evolving lithospheric architecture
through geodynamic modeling constrained by a wide range of observables from disciplines including geodesy, geology,
and geophysics.
This past year’s accomplishments include:
● Estimates of lower crustal viscosity for the Community Rheology Model.
● New models of slip rate distributions through complex fault zones including the San Gorgonio Pass area and
the San Jacinto Fault zone.
● Thin-shell finite element calculations of fault slip rates in California.
● New geodetic data and deformation models providing constraints on present-day deformation in the Western
Transverse Ranges.
● Calculations of annual loading of central and southern California faults due to the seasonal water mass cycle
in the San Joaquin Valley.
● Rate-state friction models of transition-zone seismicity along the Anza segment of the San Andreas Fault.
● Newly computed map of surface strain rate variations from 1998-2016 based on continuous GPS records.
8.1 Lower crustal viscosity
W. Shinevar (MIT), M. Behn (WHOI), G. Hirth
(Brown) and O. Jagoutz (MIT) use seismic
velocity (P and S wave) from the SCEC CVM
to constrain lower crustal viscosity. Their
analysis follows a three-step approach. First,
they use the Gibbs free energy minimization
routine Perple_X to calculate equilibrium
mineral assemblages and seismic velocities
for a global compilation of lower crustal rocks
(Hacker et al. AREPS, 2015) at various
pressures and temperatures. Second, they
use the Huet et al. (JGR, 2014) mixing model
and single-phase flow laws for major crust-
forming minerals to calculate bulk viscosity
for the predicted equilibrium mineral
assemblages. Finally, they linearly fit the
viscosity calculations to the seismic velocity
calculations. This method provides a strong
(R2>0.9) fit in the alpha-quartz regime.
Figure 8.1 shows a viscosity map for 25km
depth assuming T=700°C and έ=10-14 s-1 as
well as the viscosity-velocity data shown in
the upper right hand corner. High viscosity in
the Salton Trough is measuring mantle, rather than crustal behavior, owing to shallow Moho. The colored lines are the
estimated fit.
8.2. Model estimates of fault slip rates
P. Resor (Wesleyan), M. Cooke (U. Mass.), S. Marshall (Appalachian State), and E. Madden (U. Mass) used
mechanical models to investigate how releasing steps may affect estimates of fault slip over geologic time scales. A
suite of 2D models of idealized fault systems reveals how fault length, friction and step geometry affect kinematic
efficiency of the system and the distribution of slip rate along the faults. Although systems with longer fault segments
WELCOME
2016 SCEC Annual Meeting page 51
are more efficient, point sampling of these systems during geologic studies is unlikely to yield representative slip rate
estimates. Systems with short segments, although much less efficient, are more likely to yield representative slip rate
estimates, particularly when slip on overlapping segments is summed. A 3D model of the San Jacinto fault (Figure 8.2)
illustrates how fault system geometry may impact slip rate estimates along a real fault system. Modeled slip rates are
faster in the middle of fault segments and slower within releasing steps, consistent with published geologic estimates.
The mean slip rate along the modeled fault trace, however, is significantly slower than the simple average of the
geological rates. These results suggest that the location of geologic slip rate studies may govern their suitability for
hazard estimates and that models can be used to put point measurements of slip into the context of slip distribution
throughout a fault system.
P. Bird (UCLA) constructed dynamic
physics-based Shells Finite Element
models for southern California that
build on crustal thickness, heat-flow,
lab rheologies, fault traces and dips
and the momentum equation to
predict quasi-static stresses, strain-
rates and fault slip-rates (Figure
8.3). Bird adjusted friction on faults
to tune models to match long-term
fault slip rates from the NeoKinema
deformation model that was used in
UCERF3 and NSHM2014. The
Shells dynamic solution is revised
100 times, and in each revision the
effective friction of each of the 1000
fault elements is adjusted up or
down by a small amount (no more
than ±0.01) based on the sense of
the current slip rate error for that
fault. Results show that over 60% of
the fault elements (including most of
the San Andreas and Garlock faults)
decrease down to the lower
effective-friction limit of 0.01 and remain. Overall RMS error in fault slip rates falls from 5.3 to 1.6 mm/yr. Fault rakes
were not imposed for this study, but have excellent realism. One physical interpretation is that over 60% of active fault
area in southern California experiences near-total stress drop in large earthquakes due to dynamic weakening
mechanisms like thermal pressurization or thermal decomposition or fault melting. Thus, the peak pre-seismic shear
traction on most of these faults is not much more than the coseismic stress drop. Yet, other areas on the fault network
retain higher effective friction, concentrate shear tractions, and may serve as nucleation sites.
8.3. Western Transverse Range Deformation
In 2012, SCEC established the Ventura Special Fault Study Area (SFSA) largely based on recent work suggesting the
potential for M7.5-8.0 earthquakes in the region (Rockwell 2011, Hubbard et al. 2014, McAuliffe et al. 2015, Rockwell
et al. in review). S. Marshall (Appalachian State), G.Funning (UCR), and S. Owen (JPL) have produced a refined 3D
model based on the SCEC CFM5.0, including two potential geometries for the Ventura fault. One representation
contains a flat/horizontal ramp section (Hubbard et al., 2014), the other uses a constant dip angle (Kammerling et al.,
2003). Mechanical models of the regional faults predict no significant geologic/geodetic slip rate discrepancies, and in
general, the CFM5.0 appears to perform better than previous versions. Both representations of the Ventura-Pitas Point
fault produce model-calculated slip rates that generally fit the existing regional data; however, the two models produce
significantly different slip rates on some of the regional faults. Marshall, Funning, and Owen have processed GPS
velocities at 347 total sites throughout southern California and also processed InSAR data from the Envisat and ERS
satellites and produced two dense persistent scatterer datasets for the region. The GPS and InSAR data both show
WELCOME
2016 SCEC Annual Meeting page 52
deformation north of the Ventura fault that is consistent with interseismic deformation on a shallow dipping fault (similar
to the flat ramp model).
8.4. Transient Deformation from GPS data
R. McCaffrey (Portland State) used continuous and survey-mode GPS time series in southern California to extract
linear velocities at sites, along with parameters that describe co-seismic slip and afterslip of recent earthquakes and
volcanic deformation (1992-2016). McCaffrey also developed steady-state and time-dependent deformation models for
all California. More than 2000 time series were inverted simultaneously to estimate the source parameters for 13
earthquakes (afterslip for 10 of them), along with
volcanic deformation at Long Valley and Coso. The
linear velocities estimated were then used to
generate a steady strain rate map for California,
while the transient time histories provide a time-
dependent strain rate map.
8.5. Friction Models of Anza Seismicity
Observations from the Anza segment of the San
Jacinto Fault show that seismicity extends to depth
of 14-18 km, while the geodetically determined
locking depth is only 10 km. J. Jiang and Y. Fialko
(UCSD) developed a rate-and-state fault model
with stochastic heterogeneous frictional properties
in the transition zone to explain the difference
between the seismic and geodetic locking depths.
Their model reproduces the observed depth
relation between seismicity and geodetic locking,
with implications for potential large earthquake
behavior and interseismic phenomena for the Anza
segment.
8.6. Select Publications
● Fialko, Y., Borsa, A., Shearer P., and
Trugman, D. (2016). A comparison of
long-term changes in seismicity at The
Geysers, Salton Sea, and Coso
geothermal fields. Journal of Geophysical
Research: Solid Earth, 121(1), 225-247.
doi: 10.1002/2015JB012510.
● Field, E. H., Biasi, G. P., Bird, P., Dawson,
T., Felzer, K. R., Jackson, D. D., Johnson,
K. M., Jordan, T. H., Madden, C. L.,
Michael, A. J., Milner, K. R., Page, M. T.,
Parsons, T., Powers, P. M., Shaw, B. E., Thatcher, W. R., Weldon, R. J., & Zeng, Y. (2015). Long‐Term Time‐
Dependent Probabilities for the Third Uniform California Earthquake Rupture Forecast (UCERF3). Bulletin of
the Seismological Society of America, 105(2A), 511-543. doi: 10.1785/0120140093.
● Hearn, E., and Thatcher, W. (2015). Reconciling viscoelastic models of postseismic and interseismic
deformation: Effects of viscous shear zones and finite-length ruptures. Journal of Geophysical Research,
120(4), 2794-2819.
● Howell, S., Smith-Konter, B., Frazer, N., Tong, X., & Sandwell, D. (2016). The vertical fingerprint of earthquake
cycle loading in southern California. Nature Geoscience, 9(8), 611-614.
● Lindsey, E., and Fiako, Y. (2016). Geodetic constraints on frictional properties and earthquake hazard in the
Imperial Valley, southern California. Journal of Geophysical Research: Solid Earth,. doi:
10.1002/2015JB012516.
WELCOME
2016 SCEC Annual Meeting page 53
● Liu, S., Shen, Z., and Burgmann, R. (2015). Recovery of secular deformation field of Mojave Shear Zone in
Southern California from historical terrestrial and GPS measurements. Journal of Geophysical Research: Solid
Earth, 120(5), 3965-3990. doi: 10.1002/2015JB011941.
● Luttrell, K. M., & Smith-Konter, B. R. (2016). Limits on crustal stress in southern California from topography
and earthquake focal mechanisms. Geophysical Research Letters, (submitted).
● Williams, C. A., & Wallace, L. M. (2015). Effects of material property variations on slip estimates for subduction
interface slow slip events. Geophysical Research Letters, (accepted).
WELCOME
2016 SCEC Annual Meeting page 54
9. Earthquake Forecasting and Predictability (EFP)
The Earthquake Forecasting and Predictability (EFP) group facilitates a range of studies aimed at improved data and
methods for developing earthquake forecasting techniques and assessing earthquake predictability.
This past year’s accomplishments include:
● Improvement of forecast and hazard models of seismicity, with focus on seasonal (hydrospheric, temperature)
loadings and moment deficit rate estimations.
● Characterization of fluid-induced earthquake sequences, due to both natural and man-made fluid sources.
● Constraints on stress field heterogeneity through borehole measurements, earthquake stress drop estimates,
and shear-wave splitting observations.
● Further improvement of data-center operations and processing of the waveform archive to produce the high-
quality earthquake catalogs for southern California that are essential for hypothesis generation and testing.
● Improved methods of earthquake cluster identification, with applications to discriminating between tectonic
and human-induced seismicity and swarm analysis.
9.1 Earthquake Forecast and Hazard Studies
An ongoing collaboration among
academia, government, private sector
and CSEP is aimed at constructing and
testing a global earthquake rate model
as a function of magnitude at 0.1 by 0.1
degree resolution (Bird et al., 2015).
Nicknamed “GEAR1”, the model can be
used by the Global Earthquake Model
(GEM) and others in seismic hazard
estimation. GEAR1 covers magnitudes
5.8 and larger and time intervals from
years to decades with no explicit time-
dependence. During this year, a group
led by Kagan and Jackson obtained
new results in the statistical analysis of
global earthquake catalogs with special
attention to the largest earthquakes,
and examined the statistical behavior of
earthquake rate variations. These
results serve as an input for updating a recent earthquake forecast based on past earthquakes and geodetic strain
rates. The interseismic Moment Deficit Rate (MDR) is crucial input for earthquake hazard estimates that can be
determined geodetically using various inverse methods. Segall and Johnson presented and evaluated a number of
methods for estimating the MDR using fault-specific models. The estimates of the MDR in the Southern California fault
system range from (1.7–2.0)x10^19 Nm/yr at the 95% confidence level. This is equivalent to 150-year earthquake
moment magnitudes of 8.2-8.3, resulting in nearly one magnitude 8 earthquake of unaccounted-for moment deficit
accumulation since the 1857 Fort Tejon earthquake (Figure 9.1).
Hydrospheric loads, pore pressure fluctuations, and temperature gradients cause seasonal crustal stress variations. In
California, these processes follow an annual periodic cycle and each contributes to the crustal deformation. A project
led by Burgmann, Johnson and Dutilleul tested the sensitivity of earthquake occurrence with respect to small external
stressing due to hydrospheric loading.The results suggest a seasonally varying stress on the central SAF with a peak-
to-peak amplitude of ~0.7 kPa and a larger sensitivity of dipping faults in the Coast Ranges to the changing loads
(Dutilleul et al., 2016).
9.2 Fluid-Induced Earthquakes
Fluid injection is well established as a major driver of human induced earthquakes. Both the central U.S. and California
experienced a rapid increase in waste water injection activity since 2001, posing a substantial risk for near-by urban
centers. Understanding the mechanisms by which fluid pressure changes induce earthquakes may illuminate the
WELCOME
2016 SCEC Annual Meeting page 55
physics of triggering and lead to better forecasts of both human-induced and natural earthquakes. Goebel et al. (2016b)
used mapped fault structures and seismicity-data to create flow models that constrain the extent and amplitudes of
injection-induced pore-pressure perturbations. Using a range of realistic damage zone widths permeability values the
authors showed that induced pore-pressure changes can result in seismic activity at up to ~10 km distance from
injection wells. Cheng and Chen (2015) documented the characteristics of presumably-induced microseismicity directly
beneath a group of injection wells in the Salton Sea. These events are characterized by high b-value, lack of large
earthquakes, and short-duration, mixed swarm/ aftershock-type clustering. Areas farther from injection wells tend to
have low b-value, frequent larger earthquakes, and distinct swarm and aftershock-type clusters.
Hauksson et al. (2016) found that the 2015 Fillmore swarm exhibited unusual episodic spatial and temporal migrations.
The spatial-temporal seismicity patterns are best explained by fluid released by metamorphic dehydration processes,
migration of fluids along a detachment zone, and cascading asperity failures. McClure and colleagues developed a
statistical method for identifying induced seismicity from large
datasets and applied the method to decades of wastewater
disposal and seismicity data in California and Oklahoma. In
Oklahoma, the analysis finds with extremely high confidence
that seismicity associated with wastewater disposal has
occurred. In California, the analysis finds that seismicity
associated with wastewater disposal has probably occurred,
but the result is not strong enough to be conclusive.
9.3 Stress and Earthquake Stress Triggering
One path to studying the predictability of earthquakes is to
better understand the relationship between earthquake
occurrence and stress, including the background stress and
static and dynamic stress changes from natural and human-
made sources. Persaud et al. (2015) observed principal stress
directions in the LA Basin, using borehole breakouts obtained
from well logs to depths of 3 km. The found considerable
variations in stress directions related to the proximity to faults,
possibly controlled by fault geometry. These results indicate the
likely complexity that may be found in stress fields near other
active faults.
Goebel et al. (2016a) studied the stress state of the SCEC
Special Fault Study Areas using the stress drops of many small
earthquakes. They found that the San Gorgonio Pass exhibits
more variable and systematically higher stress drops than does
the Ventura Basin (Figure 9.2). This implies the general degree
of stress field heterogeneity and strain localization may
influence stress drops more strongly than ambient stress
levels, with more localized faulting and homogeneous stress
fields favoring lower stress drops. Li et al (2015) investigated
shear wave splitting near the San Jacinto Fault Zone and found
fast directions in some areas consistent overall with the
maximum horizontal compression direction, and in others likely reflecting fault zone damage. They found no significant
temporal changes of the fast direction and delay times within the 2-year study period.
9.4 High-quality Data
High-quality seismological data is essential for hypothesis generation and hypothesis testing. SCEC continues to
support the catalog archived by the SCEDC, which is the most complete archive of seismic data for any region in the
United States and has contributed significantly to numerous scientific papers addressing earthquake predictability. A
continuing collaboration between Caltech and UCSD has focused on improved earthquake locations and focal
mechanisms using waveform cross-correlation and S/P amplitude ratios. The catalogs from this collaboration (e.g.
Hauksson et al., 2012; Yang et al, 2012) have been widely used, leading to new results on a number of topics relevant
WELCOME
2016 SCEC Annual Meeting page 56
to better understanding earthquake predictability, including earthquake triggering, foreshock patterns, swarm
characteristics, state of stress, and earthquake scaling, which would not have been possible with standard catalogs.
9.5 Earthquake Simulators
Earthquake simulators are a class of computational simulations that attempt to mirror the topological complexity of fault
systems on which earthquakes occur. In addition, simulators can account for the physics of friction and elastic
interactions between fault elements. Wilson et al. (2016) applied spatial verification methods similar to those used in
the RELM study to a number of earthquake simulators. In order to compare the simulator outputs with real earthquakes,
which often occur off of known faults, two methods were introduced to redistribute the seismicity of simulated and
observed quakes to the same regions. The first is a simplistic approach, whereby observed earthquakes are shifted to
the nearest fault element. The second method is a variation of the Epidemic type aftershock (ETAS) model, which
distributes the simulator catalog seismicity over the entire test region. It was found that the nearest-neighbor mapping
produced poor forecasts, while the modified ETAS method produced rate maps that agreed with observations.
9.6 Earthquake Clustering
SCEC-funded research resulted in greatly improved earthquake locations, focal mechanisms, and estimates of stress
drop for southern California earthquakes. These products are now used to clarify physical processes associated with
faulting in the crust and address a number of issues related to seismic hazard. Zhang and Shearer (2016) developed
a new method of swarm identification and performed a systematic analysis of swarms in southern California. The results
show that swarms are heterogeneously distributed in time and space and are likely related to foreshock sequences in
some regions. Zaliapin and Ben-Zion (2015) documented and quantified effects of catalog uncertainties on results of
statistical cluster analyses of seismicity in southern California. Zaliapin and Ben-Zion (2016) analyzed statistical
features of background and clustered subpopulations of earthquakes in different regions in an effort to distinguish
between human-induced and natural seismicity. Induced seismicity is shown to have (i) higher rate of background
events, (ii) faster temporal offspring decay, (iii) higher rate of repeating events, (iv) larger proportion of small clusters,
and (v) larger spatial separation between parent and offspring. The results can inform a range of studies focused on
small-magnitude seismicity patterns in the presence of cat alog uncertainties, as well as to improve seismic hazards
assessment related to induced earthquakes.
9.7 Select Publications
● Bird, P., D. D. Jackson, Y. Y. Kagan, C. Kreemer, and R. S. Stein (2015). GEAR1: a Global Earthquake Activity
Rate model constructed from geodetic strain rates and smoothed seismicity, Bull. Seismol. Soc. Amer., 105
(5), 2538-2554, doi:10.1785/0120150058, (plus electronic supplement). Publication 2075, SCEC
● Cheng, Y., and X. Chen (2015), Effect of geothermal operations on seismic characteristics in the Salton Sea
geothermal field, SEG Tech. Progr. Expand. Abstr., 5063–5068, doi:10.1190/segam2015-5882527.1.
● Dutilleul, P., C. W. Johnson, R. Bürgmann, Y. Wan, and Z.-K. Shen (2015), Multifrequential periodogram
analysis of earthquake occurrence: An alternative approach to the Schuster spectrum, with two examples in
central California, Journal of Geophysical Research: Solid Earth, 120(12), 8494-8515. SCEC contribution
number 6160.
● Goebel, T. H. W., E. Hauksson, A. Plesch, and J. Shaw (2016a) Detecting significant stress drop variations in
large micro-earthquakes datasets: A comparison between a convergent step-over in the San Andreas Fault
and the Ventura thrust fault system, southern California; Special Volume of Pure and Applied Geophysics;
submitted, January 2016.
● Goebel T.H.W., S.M. Hosseini, F. Cappa, E. Hauksson, J.-P. Ampuero, F. Aminzadeh & J.B. Saleeby (2016b).
“Wastewater disposal and earthquake swarm activity at the southern end of the Central Valley, California”,
Geophys. Res. Lett., 43, 1-8, doi:10.1002/2015GL066948.
● Li, Z., Z. Peng, Y. Ben-Zion, and F. Vernon (2015), Spatial variations of shear-wave anisotropy near the San
Jacinto Fault Zone in southern California, J. Geophys. Res. Solid Earth, 120, 8334-8347, doi:
10.1002/2015JB012483.
● Hauksson, E., J. Andrews, A. Plesch, and J. H. Shaw, D. R. Shelly, The 2015 Earthquake Swarm Near
Fillmore, California: Possible Dehydration Event Near the Bottom of the Over-Pressurized Ventura Basin, in
review, Seismol. Res. Letters, Feb., 2016.
WELCOME
2016 SCEC Annual Meeting page 57
● Hauksson, E. and W. Yang, and P. M. Shearer (2012), Waveform Relocated Earthquake Catalog for Southern
California (1981 to June 2011); Bull. Seismol. Soc. Am., 102, no. 5, 2239–2244, doi: 10.1785/0120120010.
● Persaud, P., J. M. Stock, and D. Smith (2015), Evidence of sub Kilometer-scale Variability in Stress Directions
near Active Faults: An Example from the Newport-Inglewood Fault, Southern California, American
Geophysical Union Fall Meeting, 2015, Abstract T23C-2972.
● Wilson, J. M., M. R. Yoder, J. B. Rundle, D. L. Turcotte, and K. W. Schultz (2016) "Spatial Evaluation and
Verification of Earthquake Simulators, Pure Appl. Geophys., to appear.
● Yang, W., E. Hauksson and P. M. Shearer (2012), Computing a large refined catalog of focal mechanisms for
southern California (1981–2010): Temporal stability of the style of faulting, Bull. Seismol. Soc. Am., 102, 1179–
1194, doi: 10.1785/0120110311.
● Zaliapin, I. and Y. Ben-Zion (2015) Artifacts of earthquake location errors and short-term incompleteness on
seismicity clusters in southern California. Geophys. J. Intl., 202 (3): 1949-1968. doi: 10.1093/gji/ggv259
● Zaliapin, I. and Y. Ben-Zion (2016) Discriminating characteristics of tectonic and human-induced seismicity.
Bull. Seismol. Soc. Am., 106(3), 846-859, doi: 10.1785/0120150211,
● Zhang, Q., and P. M. Shearer (2016) A new method to identify earthquake swarms applied to seismicity near
the San Jacinto Fault, California, Geophys. J. Int., doi: 10.1093/gji/ggw073, 2016.
WELCOME
2016 SCEC Annual Meeting page 58
10. Ground Motion Prediction (GMP)
The primary goal of the Ground-Motion Prediction interdisciplinary working group is to develop and implement physics-
based simulation methodologies that can predict earthquake strong-motion waveforms over the frequency range 0-10
Hz and beyond. Both source and media characterization play a vital role in ground-motion prediction and are important
areas of research for the group. Note that GMP focus group also shares interests with the Earthquake Engineering
Implementation Interface (EEII) and Special Projects, and consult those sections for additional GMP-related research.
This past year’s accomplishments include:
● Anderson and co-workers, and Archuleta and Crempien introduced multi-segment ruptures into their
respective Broadband Platform modules.
● Asimaki et al. developed a model that translates Vs30 into a generic velocity profile suitable for use in wave
propagation-based ground motion models.
● Barbour and Baltay compare direct and inferred strains from all M ≥ 4.5 earthquakes on the San Jacinto Fault
since the Plate Boundary Observatory borehole network was installed (5 events), finding that measured strains
are comparable to estimates from PGV (peak ground velocity) scaled by S-wave slowness. We compare the
strains and converted PGV values for these events to NGA-West2 ground-motion prediction equations
(GMPEs), also scaled by the S-wave slowness. While there is an overall match between the data and GMPEs,
we observe a significant departure from expected attenuation behavior: peak inferred strains in the Los
Angeles basin resulting from moderate earthquakes in the Anza region are significantly larger than predicted
by GMPEs, and cannot be understood in terms of site-to-site variations in Vs30 alone.
● Bydlon and Dunham studied ground motions in areas experiencing induced seismicity. To aid the effort of
constraining near-source GMPEs associated with induced seismicity, they integrate synthetic ground motion
data from simulated earthquakes into the process. Preliminary results indicate that ground motions can be
amplified if the source is located in the shallow, sedimentary sequence compared to the basement. Source
depth could therefore be an important variable to define explicitly in GMPEs instead of being incorporated into
traditional distance metrics.
● Moschetti et al. have generated a database of more than 160,000 records from about 3200 induced
earthquakes in Oklahoma and Kansas. They measure and compile the strong ground motions and compare
these average ground motion intensity measures (IMs) with existing ground motion prediction equations
(GMPEs). Mean IMs exhibit a clear break in geometrical attenuation at epicentral distances of about 50–70
km, which is consistent with previous studies in the CEUS. At close distances, the observed IMs are lower
than the predictions of the seed GMPEs of the NGA-East project (and about consistent with NGA-West-2
ground motions). This ground motion database may be used to inform future seismic hazard forecast models
and in the development of regionally appropriate GMPEs.
● Olsen and Takedatsu have started preparing the SDSU BBP module for the next generation seismic hazard
analysis. They find reasonable fits between durations from simulations and data, which however depend
strongly on the specifics of how the values are calculated. In general, intra-event standard deviations for the
Part A PSAs generated by the SDSU module using 50 realizations are in agreement with those from data.
They also modify their module to fit median PSAs according to NGA-west2 (SCEC Validation Exercise, Part
B).
● Riano et al. have performed a series of 5-Hz 3D simulations for the region of the Oxnard plain including the
surrounding geologic structures, and focus on the effects of the topography on the ground response, by
comparing the results from models with and without the surface topography. In general, the amplitudes of the
Fourier spectra in the model with topography are larger than the those in the model for frequencies greater
than 2 Hz. Simulations are done using Hercules.
● Sleep et al. tested the prediction that strong Love waves and near-field velocity pulses damage the shallow
subsurface. After a few events, the shear modulus self organizes as a function of depth so that frictional failure
barely occurs. Broader SCEC implications are that friction is an appropriate rheology for shallow exhumed
sedimentary rocks and the major earthquakes with particle velocities of 1-2 m/s sometimes occur in the
Parkfield segment of the San Andreas Fault. Overall shallow rock is a fragile geological feature analogous to
precarious rocks.
● The SCEC High-F group has worked toward reproducing the ground motions from the 2014 M 5.1 La Habra,
California, earthquake in the greater Los Angeles region over a simulation domain of 180 km x 135 km, and
WELCOME
2016 SCEC Annual Meeting page 59
62 km in depth. The simulations are done using three high-performance computing codes; two of them
implement the finite difference method and the third one implements a finite element approach. The models
are tailored to satisfy a maximum frequency of 5 Hz and a minimum shear wave velocity of 500 m/s. The group
compared synthetics from the three codes for the point source model, using the 3D regional structure and
frequency independent attenuation (Q) models. The comparisons between the three codes exhibit very good
agreement. For the validation stage, they compared seismograms collected at 300+ recording stations from
different regional strong-motion seismic monitoring networks with synthetics obtained for the three different
source models. Initial results indicate that extended source models, even for a moderate-size earthquake like
the one considered here, tend to lead to better fits with data.
● Withers et al. (2016) generated an ensemble of 0-8 Hz rough-fault blind-thrust simulations. They found that
the similarity between the intra-event variability of the simulations and observations increases when small-
scale heterogeneity and plasticity are included. In addition to GMPEs, they compared with simple proxy
metrics to evaluate the performance of the deterministic models and to determine the importance of different
complexities within the model. They found that 3D heterogeneity, at both the long and short scale-lengths, is
necessary to agree with data, and should be included in future simulations to best model the ground motion
from earthquakes.
● Ongoing work includes a kinematic rupture generator (Savran and Olsen), 3D dynamic rupture and ground
motion modeling for the Hosgri-Shoreline fault and Ventura’s fault system.
10.1 Deterministic
nonlinear 4-Hz M7.7 San
Andreas rupture
Roten et al. simulate M 7.7
earthquakes on the southern
San Andreas fault with a spatial
resolution of 25 m, which allows
us to resolve frequencies up to 4
Hz using a minimum shear-wave
velocity of 500 m/s (see Figure
10.1). The simulations are
performed for a linear medium
and a non-linear medium
governed by Drucker-Prager
plasticity. The mesh is defined
by SCEC CVM-S4, including a
geotechnical layer, small-scale
heterogeneities and frequency-
dependent quality factors
Spectral accelerations at 2s (2s-
SAs) frequently exceed 0.5g near the fault and in the Los Angeles basin in the linear case. Nonlinearity reduces long-
period ground motions in these locations by 25 to 50%. At higher frequencies, plastic effects are most pronounced
within ~20 km of the fault, where 1s-SAs are reduced by up to 50% and 0.5s-SAs by up to 75% with respect to linear
simulations. However, accounting for plasticity brings synthetic 0.5s-SAs in agreement with ground motion prediction
equations (GMPEs) at these fault distances, while linear 0.5s-SAs exceed GMPEs by more than one standard
deviation.
WELCOME
2016 SCEC Annual Meeting page 60
10.2 High-resolution seismic imaging of
the very shallow crust
Ben-Zion and co-workers have developed and
used several techniques to derive from dense
seismic arrays (sensor spacing 10-30 m)
crossing the San Jacinto fault zones at several
locations (Figure 10.2) (i) attenuation
coefficients, (ii) seismic velocities and (iii)
anisotropy of the subsurface material. The results
indicate Q values of ~6-30 in the top ~30 m of the
crust [Liu et al., 2015] and seismic velocities of
~300-700 m/s at the top ~50-500 m [Roux et al., 2016; Hillers et al. 2016].
10.3 A stochastic Vs30-dependent velocity model of
near-surface sediments in Southern California (GTL
v2.0)
Asimaki et al. have developed and validated (in 1D) a Vs30-
dependent shear wave velocity model of the near surface soft
sediments (previously referred to as Geotechnical Layer or GTL),
based on the statistics of the AARA (Yong et al, 2013) velocity
profiles with VS30 ranging from 150 m/s to 800 m/s. They have also
developed a spatially correlated random realization algorithm using
the statistics of the AARA profiles, to populate the near surface of
the 3D UCVM domain with our VS30-dependent velocity profiles. So
far the model is implemented in 2D, with plans to extend to random
realizations to 3D. See Figure 10.3 on the right.
10.4. Select Publications
● Hillers, G., P. Roux, M. Campillo, Y. Ben-Zion (2016). Focal
spot imaging based on zero lag cross correlation amplitude
fields: Application to dense array data at the San Jacinto
fault zone, J. Geophys. Res., in review.
● Liu, X., Y. Ben-Zion and D. Zigone (2015). Extracting
seismic attenuation coefficients from cross-correlations of
ambient noise at linear triplets of stations, Geophys. J. Int., 203, 1149–1163.
● Roten, D., Y. Cui, K.B. Olsen, S. M. Day, K. Withers, W. Savran and P. Wang (2016). "High-frequency
Nonlinear Earthquake Simulations on Petascale Heterogeneous Supercomputers", SC16, November 13-18,
2016, Salt Lake City, UT, accepted.
● Roten, D., K.B. Olsen, S. M. Day and Y. Cui (2016). Quantification of fault zone plasticity effects with
spontaneous rupture Simulations, PAGEOPH, submitted.
● Roten, D., K. B. Olsen, Y. Cui, and S. M. Day (2015), Quantification of fault zone plasticity effects with
spontaneous rupture simulations, Workshop on Best Practice in Physics-Based Fault Rupture Models for
Seismic Hazard Assessment of Nuclear Installations, Vienna, Austria, Nov 18-20.
● Roux, P., L. Moreau, A. Lecointre, G. Hillers, M. Campillo, Y. Ben-Zion, D. Zigone and F. Vernon, (2016). A
methodological approach toward high-resolution surface wave imaging of the San Jacinto Fault Zone using
ambient-noise recordings at a spatially dense array, Geophys. J. Int., 206, 980–992.
● Savran, W.H., and K.B. Olsen (2016). Model for Small-Scale Crustal Heterogeneity in Los Angeles Basin
Based on Inversion of Sonic Log Data, Geophys. Jour. Int. 205, 856-863. SCEC Contribution 6178.
● Sleep, N.H., and N. Nakata (2015). Nonlinear attenuation from the interaction between different types of
seismic waves and interaction of seismic waves with shallow ambient tectonic stress, Geochemistry,
Geophysics, Geosystems Geochem. Geophys. Geosyst., 16, 2336–2363, doi:10.1002/2015GC005832.
WELCOME
2016 SCEC Annual Meeting page 61
● Sleep, N. H. (2016). Heat flow, strong near-fault seismic waves, and near fault tectonics on the central San
Andreas Fault, Geochem. Geophys. Geosyst., 17, doi:10.1002/ 2016GC006280. SCEC Contribution 6259.
● Sleep, N.H. (2016). Shallow Sedimentary Rock as a Fragile Geological Feature: Effects of Clay Content and
Hydrology on Frictional Strength, SCEC Contribution 6431.
● Withers, K.B., K.B. Olsen and S.M. Day (2015). Memory Efficient Simulation of Frequency Dependent Q, Bull,
Seis. Soc. Am. 105, 3129-3142, doi: 10.1785/0120150020. SCEC Contribution 6002.
WELCOME
2016 SCEC Annual Meeting page 62
11. Earthquake Engineering
Implementation Interface
(EEII)
The implementation of SCEC research
for practical purposes depends on
interactions with engineering
researchers and organizations, and with
practicing engineers, building officials,
insurers, emergency managers, and
other technical users of our information.
An important area of EEII work is in the
validation and utilization of ground
motion simulations. With the important
milestone of completion of the
BroadBand Platform validation effort,
there is now significant data and
computational infrastructure that is being
utilized in this area.
This past year’s accomplishments
include:
● Vetting and publication of a
range of simulation validation
metrics
● Implementation of Ground
Motion Simulation Validation
Gauntlets.
● Utilization of Ground Motion
Simulations to Produce Urban
Seismic Hazard Maps.
11.1 New metrics for ground
motion simulation validation
With the Ground Motion Simulation
Validation Technical Activity Group
having been active for a number of years,
results from that group are reaching
maturity and are being disseminated to a
broad audience in the science and
engineering communities. New results
this year include validation of simulations
using new ground motion metrics that
measure temporal and
temporal/frequency evolution of ground
motion simulations, as compared to
recordings. Afshari and Stewart (2016)
developed a new predictive model for
ground motion duration, using physically parameterized source, path and site terms. This improved model allows for
more refined comparisons of simulated durations versus observations. Rezaeian et al. (2015) evaluated new metrics
for temporal evolution and frequency nonstationarity of ground motions, using random vibrations theory tools (see
Figure 11.1).
WELCOME
2016 SCEC Annual Meeting page 63
11.2. Implementation of validation metrics in the Broadband Platform
The SCEC Ground Motion Simulation Validation (GMSV) Technical Activity Group (TAG) aims to develop and
implement, via collaboration between ground motion modelers and engineering users, testing/rating methodologies for
simulated ground motions to be used in engineering applications. At a 2014 progress and planning workshop of the
GMSV TAG, the need for a multi-PI collaborative project that builds on the knowledge from these previous projects was
discussed. SCEC has begun to fill this need by implementing a “gauntlet” of 10 previously published ground motion
validation parameters and corresponding empirical ground motion models on the Broadband Platform (BBP). The focus
was on relatively simple validation parameters (e.g., strong motion duration) and the Broadband Platform is consistent
with recommendations from the initial planning workshop of the GMSV TAG in 2011. The work has involved identifying
the most promising and mature metrics for consideration, developing robust algorithms for computing these metrics,
and implementing the metrics on the BBP so that they are computed in-line with any new simulations (Figure 11.2).
Results have been obtained and evaluated by the project PIs, the GMSV TAG, and a range of the broader engineering
seismology community.
11.3. Use of CyberShake for Los Angeles long-period ground motion hazard maps
The SCEC Committee on Utilization of Ground Motion Simulations (UGMS) aims to develop long-period response
spectral acceleration maps for the Los Angeles region for inclusion in NEHRP and ASCE 7 Seismic Provisions and in
Los Angeles City Building Code. The maps are to be based on 3-D numerical ground-motion simulations, and ground
motions computed using latest empirical ground-motion prediction equations from the PEER NGA project. The work of
the UGMS committee is being coordinated with (1) the SCEC Ground Motion Simulation Validation Technical Activity
Group (GMSV-TAG), (2) other SCEC projects, such as CyberShake and UCERF, and (3) the USGS national seismic
hazard mapping project.
Further progress this year includes advancements in quantifying and accounting for near-surface soils (the
“geotechnical layer”) in the simulations. Additionally, a plan for utilizing CyberShake predictions as part of a logic tree
with empirical ground motion predictions has been developed and refined. Finally, a USGS web app look-up tool has
been planned, to facilitate access to the results in a convenient and archival manner. And numerical results continue
to be produced and evaluated. The work has proceeded with regular meetings attended by a number of engineering
stakeholders and USGS scientists and engineers, to ensure the relevance of the produced products, and to achieve
buy-in from the ultimate users of the work.
11.4. Select Publications
● Afshari, K., and Stewart, J. P. (2016). “Physically Parameterized Prediction Equations for Significant Duration
in Active Crustal Regions.” Earthquake Spectra, in press.
● Rezaeian, S., Zhong, P., Hartzell, S., and Zareian, F. (2015). “Validation of Simulated Earthquake Ground
Motions Based on Evolution of Intensity and Frequency Content.” Bulletin of the Seismological Society of
America, 105(6), 3036–3049.
● Burks, L. S., and Baker, J. W. (2016). “A predictive model for fling-step in near-fault ground motions based on
recordings and simulations.” Soil Dynamics and Earthquake Engineering, 80(1), 119–126. SCEC contribution
number 1964.
● Burks, L. S., Zimmerman, R. B., and Baker, J. W. (2015). “Evaluation of Hybrid Broadband Ground Motion
Simulations for Response History Analysis and Design.” Earthquake Spectra, 31(3), 1691–1710. SCEC
contribution number 1794.
● Bradley, B. A., Burks, L. S., and Baker, J. W. (2015). “Ground motion selection for simulation-based seismic
hazard and structural reliability assessment.” Earthquake Engineering & Structural Dynamics, 44(13), 2321–
2340. SCEC Contribution 6089.
WELCOME
2016 SCEC Annual Meeting page 64
12. Working Group on California Earthquake Probabilities (WGCEP)
The WGCEP is a collaboration between SCEC, the USGS, and CGS aimed at developing official earthquake-rupture-
forecast models for California. The project is closely coordinated with the USGS National Seismic Hazard Mapping
Program, and has received financial support from the California Earthquake Authority (CEA) and the Keck Foundation
(via the CISM project).
Building on the time-independent and time-dependent Uniform California Earthquake Rupture Forecast models
(UCERF3-TI and UCERF3-TD), published in previous years, our recent efforts have been focused on adding a
spatiotemporal clustering component in order to represent aftershocks, induced seismicity, and otherwise triggered
events, with the ultimate goal being to serve as a basis for Operational Earthquake Forecasting (OEF). This effort was
bolstered by a workshop on the “Science of OEF” held in October 2015 at the USGS Powell Center in Fort Collins, CO.
A primary goal of this study has been to bring fault-based information into OEF, which has traditionally been based
solely on GR and Omori-Utsu statistics (Jordan et al., 2011). In fact, the California Earthquake Prediction Evaluation
Council (CEPEC) is known to convene when a magnitude ~5 earthquake occurs near the San Andreas fault (Jordan
and Jones, 2010), but not when such an earthquake occurs well away from known faults, implying that fault proximity
is an important consideration, not to mention the overall activity rate and elastic-rebound readiness of the fault.
To explicitly consider such fault-based information, we have added an Epidemic Type Aftershock Sequence (ETAS)
component to the previously published time-independent and long-term time-dependent forecasts. This combined
model, referred to as UCERF3-ETAS, collectively represents a relaxation of segmentation assumptions, the inclusion
of multi-fault ruptures, an elastic-rebound model for fault-based ruptures, and a state-of-the-art spatiotemporal
clustering component. It also represents an attempt to merge fault-based forecasts with statistical seismology models,
such that information on fault proximity, activity rate, and time since last event are considered in OEF. A paper has now
been submitted to BSSA (Field et al., 2016; accepted pending minor revisions, Figure 12.1), which describes not only
the model, but several unanticipated challenges that were encountered as well, including a need for elastic rebound
and characteristic magnitude-frequency distributions on faults, both of which are required to get realistic triggering
behavior. UCERF3-ETAS produces synthetic catalogs of M≥2.5 events, conditioned on any prior M≥2.5 events that are
input to the model. The paper shows results with respect to both long-term (1,000-year) simulations, as well as for 10-
year time periods following a variety of hypothetical scenario main shocks. While the results are very plausible, they
are not always consistent with the
simple notion that triggering
probabilities should be greater if a
main shock is located near a fault.
Important factors include whether
the magnitude-frequency
distribution near faults includes a
significant characteristic
earthquake component, as well as
whether large triggered events can
nucleate from within the rupture
zone of the main shock. Because
UCERF3-ETAS has many sources
of uncertainty, as will any
subsequent version or competing
model, potential usefulness will
need to be considered in the
context of actual applications.
12.1. Select Publications
● Field, E. H., K. R. Milner, J. L. Hardebeck, M. T. Page, N. van der Elst, T. H. Jordan, A. J. Michael, B. E. Shaw,
and M. J. Werner (2016). A Spatiotemporal Clustering Model for the Third Uniform California Earthquake
Rupture Forecast (UCERF3-ETAS) – Toward an Operational Earthquake Forecast, Bull. Seism. Soc. Am.,
Submitted.
WELCOME
2016 SCEC Annual Meeting page 65
● Jordan, T. H., Y.-T. Chen, P. Gasparini, R. Madariaga, I. Main, W. Marzocchi, G. Papadopoulos, G. Sobolev,
K. Yamaoka, and J. Zschau (2011). Operational earthquake forecasting: state of knowledge and guidelines
for implementation, final report of the International Commission on Earthquake Forecasting for Civil Protection,
Annals Geophys., 54(4), 315-391, doi:10.4401/ag-5350.
● Jordan, T. H., and L. M. Jones (2010). Operational earthquake forecasting: some thoughts on why and how,
Seism. Res. Lett. 81 (4), 571–574.
WELCOME
2016 SCEC Annual Meeting page 66
13. Collaboratory for the Study of Earthquake Predictability (CSEP)
The global Collaboratory for the Study of Earthquake Predictability (CSEP, www.cseptesting.org) supports a global
program for research into earthquake predictability. CSEP provides a cyber-infrastructure for conducting rigorous
earthquake predictability experiments in a controlled and reproducible environment.
This past year’s accomplishments include:
● Ensemble modeling capability to construct ensemble forecasts from individual models using Bayesian Model
Averaging (BMA) and other techniques. These have been implemented for the California 1-day forecast group
as well as the retrospective Canterbury experiment.
● A strengthened testing methodology that can evaluate forecasts specified as simulated catalogs that can be
arbitrarily clustered and dependent, and can include epistemic uncertainties. Target models include
simulation-based statistical clustering models and UCERF3-ETAS.
● Software efficiency enhancements to enable and start high-resolution global experiments of models such as
the Global Earthquake Activity Rate (GEAR) model developed with the Global Earthquake Model (GEM)
Foundation.
● Continued hardware and software support of CSEP activities to conduct earthquake predictability experiments
around the globe, now featuring 442 models under prospective and independent evaluation (including 6 new
models installed last year in California and New Zealand).
● The development of Python cookbooks for using MiniCSEP, a try-at-home version of CSEP software available
to download and use for research and development before submission to CSEP.
● Continued international collaboration with CSEP groups in New Zealand, Germany, Switzerland, Italy and
Japan, as well as the Global Earthquake Model (GEM) Foundation.
13.1 Bayesian Forecasting in
California
Ensemble forecasts have two advantages.
They provide a quantitative and objective
approach to constructing forecasts when
multiple models are available that cannot be
ruled out a priori. They may also result in
greater predictive skill than individual models if
individual models contain complementary
information about future earthquakes.
CSEP has implemented ensemble modeling
capabilities in two testing regions. In California,
Werner et al. (2016) implemented Bayesian
Model Averaging (BMA) techniques to
construct 1-day ensemble forecasts. Figure
13.1 illustrates the evolution of the posterior
weights of individual models within the
ensemble. In BMA, posterior weights depend
solely on each model’s individual past
performance. Initially, all models carry the
same weight in the ensemble. The log-likelihood score then determines the relative contribution of each model to the
ensemble. Figure 13.1 shows several interesting features that warrant further study. Initially, the weights of most
models decay, but intermittent spikes disrupt this pattern. A non-parametric kernel smoothing model (K3) dominates
the ensemble. In May 2014, however, an abrupt change occurs: model K3 is supplanted by a model that averages an
ETAS model and the K3 model. This change coincides with a M7 earthquake and illustrates interesting changes in the
relative performance of models over time. The evolution of posterior weights can thus reveal strengths and weaknesses
of forecast models. The BMA ensemble performed slightly worse than the best individual model. BMA techniques are
thus a viable option for Operational Earthquake Forecasting (OEF) systems when the best individual model is unknown
a priori.
WELCOME
2016 SCEC Annual Meeting page 67
13.2. Ensemble Modeling of the Canterbury, New Zealand, Earthquake Sequence
Bayesian Model Averaging (BMA) is not
the only way to construct an ensemble,
and it has been criticized from a
theoretical perspective. Taroni et al
(2016) modelled the Canterbury, New
Zealand, sequence with several
ensemble approaches to elucidate
differences. Taroni et al. implemented
four different ensemble strategies. The
strategies for calculating model weights
in the ensembles differed by how the
past performance of the models was
evaluated. The BMA method revealed
the strongest sensitivity to past
performance, while the other ensembles
were more conservative in how weights
changed. Figure 13.2 shows the weight
evolutions of four ensemble methods for
forecasts based on a real-time input catalog (left) as well as a post-processed input catalog (right). The time series
reveal complex relationships between observed seismicity and model performance over time, including abrupt changes
in dominating models.
13.3. High-Resolution Global
Experiments: Evaluating GEAR1
CSEP joined forces with the Global Earthquake
Model (GEM) Foundation for the purpose of
evaluating GEM’s recent global earthquake
forecast. Bird et al. (2015) produced GEM’s high-
resolution global long-term forecast, named the
Global Earthquake Activity Rate model version
1.0 (GEAR1). GEAR1 is now under retrospective
and prospective testing within CSEP and first
results are available. To be able to begin testing,
we needed to develop substantially more efficient
CSEP formats and testing procedures, because
of the high (0.1 by 0.1 degree) spatial resolution
of the global forecast. We implemented several
efficiencies that, in the most extreme case, cut
processing time from weeks to minutes. The two
most important changes are (i) a magic index that allows calculation (rather than a search) of the index of a bin, and
(ii) making simulations of model likelihood scores more efficient. We evaluated three global high-resolution models:
GEAR1 and two strain-rate based forecasts, namely GSRM (Global Strain Rate Map, Bird et al., 2010) and an updated
version GSRM2.1 (Bird and Kreemer, 2015). Over the prospective testing period between October 2015 and April 2016,
59 earthquakes Mw> 5.77 occurred, very close to the expected number of all three models. All models pass the space,
magnitude and conditional likelihood tests, i.e. the observations are consistent with the forecasts. In a head-to-head
comparison, shown in Figure 13.3, GEAR1 achieves an information gain of 0.5 over both strain-rate based forecasts,
indicating that integrating smoothed seismicity into strain-rate information provides greater predictive skill.
13.4. Select Publications
● Bird, P., Jackson, D. D., Kagan, Y. Y., Kreemer, C. W., & Stein, R. S. (2015). GEAR1: a Global Earthquake
Activity Rate model constructed from geodetic strain rates and smoothed seismicity. Bulletin of the
Seismological Society of America, 105(5), 2538-2554. doi: 10.1785/0120150058.
WELCOME
2016 SCEC Annual Meeting page 68
● Kagan, Y. Y., & Jackson, D. D. (2016). GEAR1 forecast: Distribution of largest earthquakes and number test,.
Geophysical Journal International, 206(1), 630-643.
● Taroni, M., Werner, M.J., Marzocchi, W., and J.D. Zechar (2016). Operational Ensemble Model Earthquake
Forecasting: Case Study of the 2010-12 Canterbury, New Zealand, Sequence. AGU Fall Meeting Abstract.
● Werner, M. J., Liukis, M., Marzocchi, W., Rhoades, D. A., Taroni, M., Zechar, Z. D., & Jordan, T. H. (2016,
08). CSEP Evaluations of 24-Hour Earthquake Forecasting Models for California: New Results and Ensemble
Models. Poster Presentation at 2016 SCEC Annual Meeting.
WELCOME
2016 SCEC Annual Meeting page 69
14. Community Modeling Environment (CME)
The Community Modeling Environment (CME) is a SCEC collaboration that develops improved ground motion forecasts
by integrating physics-based earthquake simulation software, observational data, and earth structural models using
advanced computational techniques including high performance computing. CME projects often use results, and
integrate work, from SCEC groups including Interdisciplinary Focus Groups Technical Activity Groups. The SCEC/CME
was formed in 2001 as a National Science Foundation (NSF) Information Technology Research (ITR) project, and has
worked since then as an earthquake system science research group within SCEC, that makes use of the most recent
SCEC research ideas and results, and integrates these developments into computational data products with significant
societal impact.
SCEC/CME research is organized around an iterative process of improving ground motion forecasts described in four
computational pathways, as shown in Figure 14.1. Examples of SCEC/CME research includes development of earth
structural models, curation of data sets to support forecast validation, and development of scientific software that
simulates physical processes in the earth including dynamic ruptures, and wave propagation simulations. SCEC/CME
researchers are developing ground motion simulations that produce broadband seismograms. These simulation tools
include rupture generators, low
frequency wave propagation models,
high frequency stochastic models,
non-linear site response modules, and
validation capabilities including
assembled observational strong
motion data sets and waveform-
matching goodness of fit algorithms
and information displays. Ground
motion simulation validation
computational and organizational tools
are used to establish repeatable
validation of ground motion
simulations to engineering standards.
SCEC/CME researchers are also
working to improve probabilistic
seismic hazard calculations which
require a high resolution 3D velocity
model for California, a pseudo-
dynamic rupture generator capable of generating an extended earthquake rupture forecast from UCERF3.0, highly
efficient reciprocity-based seismogram calculations, and probabilistic hazard model information system providing
access to calculation results.
SCEC/CME researchers have developed a collection, an ecosystem, if you will, of interrelated open-source seismic
hazard software distributions, each of which implements some required aspect of complex seismic hazard calculations
including OpenSHA, Broadband Platform, Unified Community Velocity Model, CyberShake, AWP-ODC, Hercules, and
others.
This past year’s accomplishments include:
● Extended the OpenSHA software to implement ETAS methods in support of operational earthquake
forecasting developments by CISM and CSEP.
● Released updated southern California velocity models including a recently developed tomography-based
central California velocity model (CCA) that includes the Santa Barbara and Bakersfield areas.
● Integrated new ground motion simulation methods into SCEC Broadband Platform computational software
and released new public versions of the BBP software.
● Performed SCEC computational research using the national largest open-science supercomputers including
ALCF Mira, OLCF Titan, NCSA Blue Waters, and TACC Stampede.
WELCOME
2016 SCEC Annual Meeting page 70
● Developed standardized seismogram processing system used on the High-F and Broadband Platform projects
to compare simulation results against each other and against observations, and to produce goodness of fit
measurements for simulations.
● Produced visualizations of ground motion simulation results used by NSF and national HPC centers as
examples of important computational research results.
● Developed high-performance scientific workflows tools, in collaboration with ISI computer scientists, that
accelerate SCEC research computing within NSF and DOE HPC computing environments.
● Evaluated the impact of alternative southern California velocity models on standard seismological goodness
of fit measures by comparing simulation results against observed seismograms.
● Worked with Civil Engineering groups to calculate site specific, risk-targeted, maximum considered
earthquake (MCER) response spectra for selected sites in Los Angeles region using CyberShake results.
14.1 Ground Motion
Validation using Alternative
Velocity Models
A team led by Ricardo Taborda at the
University of Memphis used Blue
Waters to evaluate four existing
southern California velocity models
by assessing how well each velocity
model predicted ground motion in the
greater Los Angeles region when
used as inputs to deterministic wave
propagation simulations. These
evaluations were carried out by
running multiple earthquake
simulations and then using
quantitative comparisons between
simulated motions and a collection of
recorded event data. The team used Blue Waters to simulate earthquakes within a simulation domain with a surface
area of 180 km x 135 km. Each earthquake was modeled as a point source with rupture parameters scaled according
to magnitude. The group used Hercules—a finite-element software developed by SCEC-affiliated scientists—to
simulate the ground motions for each earthquake and velocity model combination. Hercules has been used before in
multiple verification and validation exercises, and has shown to be a reliable tool for 3D earthquake ground motion
simulation. The group simulated 30 well-recorded moderate-magnitude earthquakes (3.5 to 5.5) and compared
synthetics with data recorded by seismic networks on over 800 stations. Each of the 120 simulations (30 earthquakes,
4 velocity models) was run with a maximum frequency of 1 Hz and a minimum shear wave velocity of 200 m/s. The
comparisons between data and synthetics were ranked quantitatively by means of standard seismological goodness-
of-fit (GOF) criteria. The group analyzed the regional distribution of the GOF results for all events and all models, and
ranked the performance of each velocity
model (Figure 14.2) and were able to
identify that one of the southern California
velocity models yields consistently better
results (Taborda et al., 2016).
14.2. Ground Motion Simulation
Validation at 4Hz
The SCEC/CME ground motion modelers
are collaborating on 0-4Hz high-frequency
ground motion simulation validation study
using observations from the La Habra
earthquake near Los Angeles. The group
has reviewed 0-4Hz simulation results
WELCOME
2016 SCEC Annual Meeting page 71
against observations, using multiple codes, including AWP-ODC, Hercules, and RWG codes. This activity involves
running high frequency ground motion simulations using multiple codes, comparing simulation results between
methods, and comparing simulation results against observed data. Example comparisons between three computational
methods and against data, at 0.5Hz and at 4Hz are shown in Figure 14.3. As we are building the high-frequency ground
motion simulation evaluation process, more 0-4Hz validation simulations are needed, adding more complex physics
into the simulations..
14.3. High Performance Wave Propagation Software Development
The SCEC PAID Project science team, led by Yifeng Cui at SDSC, worked with the computer science from NCSA, to
optimize the nonlinear AWP-ODC GPU code for scalability and efficiency on Blue Waters. The team, in collaboration
with Kim Olsen’s group at SDSU, integrated several improvements into the AWP numerical models including support
for plasticity yielding and frequency-dependent attenuation. They also optimized code performance through yield factor
interpolation, memory tuning to increase occupancy, communication overlap using multi-streaming, and parallel I/O to
support concurrent source inputs. Simulation of non-linear material behavior requires the addition of 17 new variables
compared to linear computations, posing challenges in terms of solution-time and memory management. The improved
nonlinear code has proven to be highly scalable and efficient and achieved better overall peak performance than the
linear code despite the additional required variables and processing. Kim Olsen’s team then used the improved software
on Blue Waters to perform 0–4 Hz nonlinear ShakeOut scenario earthquake simulations. These results represent a
significant advance in our ability to perform earthquake simulations at frequencies up to 4 Hz because our codes now
include the advanced physics, including small-scale complexity of the source, nonlinear effects, and frequency-
dependent inelastic attenuation that are needed to accurately simulate these higher frequencies ground motions.
14.4. Select Publications
● Deelman, E., Vahi, K., Juve, G., Rynge, M., Callaghan, S., Maechling, P. J., Mayani, R., Chen, W., da Silva,
R., Livny, M., & Wenger, K. (2014). Pegasus, a Workflow Management System for Large-Scale Science.
Future Generation Computer Systems, Volume 46, May 2015, Pages 17–35 doi:10.1016/j.future.2014.10.008
● M. P. Moschetti, N. Luco, A. Baltay Sundstrom, O. Boyd, A. Frankel, R. Graves, M. Petersen, E. Thompson,
T. Jordan, S. Callaghan, C. Goulet, K. Milner, P. Maechling , INCORPORATING LONG-PERIOD (T>1 S)
Ground Motions from 3-D Simulations in the U.S. National Seismic Hazard Model, 16th World Conference on
Earthquake, 16WCEE 2017 Santiago Chile, January 9th to 13th 2017
● Roten, D., Y. Cui, K.B. Olsen, S. M. Day, K. Withers, W. Savran and P. Wang (2016). "High-frequency
Nonlinear Earthquake Simulations on Petascale Heterogeneous Supercomputers", SC16, November 13-18,
2016, Salt Lake City, UT, accepted.
● Roten, D., K.B. Olsen, S. M. Day and Y. Cui (2016). Quantification of fault zone plasticity effects with
spontaneous rupture Simulations, PAGEOPH, submitted.
● Goulet, C.A., Crouse, C. B. Jordan, T. H., Graves, R. Olsen, K. B., Milner, K., Silva, F., Cui, Y., Maechling, P.
J. “The SCEC Broadband Platform And Cybershake: Comprehensive Earthquake Ground Motion Simulation
Tools For Engineering Applications”, with Earthquake Engineering Research Institute (EERI) Annual Meeting,
San Francisco, CA, April 4-9, 2016,
● Goulet, C.A., Maechling, P. J., Jordan, T. H., Luco, N., Rezaeian, S. (2016) “Earthquake Ground Motion
Simulations: Scec Community Code Development And Validation Efforts”, Seismological Society of America
Annual Meeting, Reno, NV. April 20-23, 2016.
● Roten, D., Olsen, K. B., Day, S. M., Cui, Y. (2016) “Fault Zone Plasticity Effects Quantified By Spontaneous
Rupture Simulations”, Seism. Res. Lett. 87, 2B, 502-503.
● Withers, K. B., Olsen, K. B., Shi, Z., Day, S. (2016) “Ground Motion Variability From 3-D Deterministic
Broadband (0-8 Hz) Ensemble Simulations Of Mw 7.1 Strike-Slip And Mw 6.7 Blind Thrust Events
Incorporating Rough Fault Topography, Frequency-Dependent Viscoelasticity, Small-Scale Media
Heterogeneity, And Plasticity” , Seism. Res. Lett. 87, 2B, 455.
● Animation based on Titan results: https://www.youtube.com/watch?v=znWqQ7wa8-w
● Future Directions for NSF Advanced Computing Infrastructure to Support U.S. Science and Engineering in
2017-2020, Committee on Future Directions for NSF Advanced Computing Infrastructure to Support U.S.
Science in 2017-2020; (Pg 54) Computer Science and Telecommunications Board; Division on Engineering
and Physical Sciences; National Academies of Sciences, Engineering, and Medicine, ISBN 978-0-309-38961-
7 | DOI: 10.17226/21886.
WELCOME
2016 SCEC Annual Meeting page 72
15. Collaboratory for Interseismic Simulation and Modeling (CISM)
The Collaboratory for Interseismic Simulation and Modeling (CISM) is an effort to forge physics-based models into
comprehensive earthquake forecasts using California as its primary test bed. Short-term forecasts of seismic
sequences, in combination with consistent long-term forecasts, are critical for reducing risks and enhancing
preparedness. CISM seeks to improve predictability by combining rupture simulators that account for the physics of
rupture nucleation and stress transfer with ground-motion simulators that account for wave excitation and propagation.
CISM forecasting models will be tested against observed earthquake behaviors within the existing Collaboratory for the
Study of Earthquake Predictability (CSEP).
We hosted a large public project kick-off meeting at the 2015 SCEC Annual Meeting. The objectives of this meeting
were to further refine the objectives developed for the project and to encourage SCEC scientists to participate in CISM
activities. Over 30 SCEC scientists participated with presentations and in discussions on the important issues to be
addressed by the project. The main outcomes of the workshop were the confirmation of RSQSim as the main
earthquake simulator for the package, which was going to be used in conjunction with the Uniform California Rupture
Forecast 3 (UCERF3) model for the development of long-term earthquake catalogs (Dieterich and Richards-Dinger,
2010; Richards-Dinger and Dieterich, 2012; Field et al., 2013). Refinements to the tasks priorities were also proposed.
The creation of topic-targeted working groups was also explored during the workshop, with key leaders identified by
the SCEC CISM management team.
This past year’s accomplishments include:
● Documentation of the Epidemic-Type Aftershock Sequence (ETAS) version of the third Uniform California
Earthquake Forecast (UCERF3) model (see Section 12).
● Significant progress to RSQSim development and calibration against datasets and UCERF3.
● RSQSim merge to NCSA Blue Waters and parameter space exploration.
● Integration of RSQSim catalogs into OpenSHA.
● Development of visualization tools for UCERF3 and RSQSim earthquake catalogs.
● RSQSim code performance optimization.
● Use of RSQSim and High-Performance Computing in Undergraduate Students Training.
15.1 RSQSim Development and Calibration
We developed a conceptual outline for, and began tests of, a method for physics-based estimation of earthquake
probabilities in California. Current statistics-based methods for earthquake probabilities are fundamentally limited by
the inherent limitations in the observational record – inter-event times for recurrence of damaging earthquakes are
typically hundreds to thousands of years, while the instrumental observation record is on the order of 100 years. Hence,
site-specific statistics are severely limited, and the probability density distributions for recurrence of earthquakes
(required for analysis of earthquake probabilities) consist of hypothetical density functions that are partially constrained
by paleoseismic investigations of prehistoric earthquakes at scattered sites, which have uncertain dates and
magnitudes.
The physics-based method uses simulated California seismicity catalogs generated by RSQSim to define a) empirical
site-specific probability density distributions for recurrence of slip, and b) earthquake source characteristics (extent of
earthquake rupture, slip, and source time functions). Inputs to the simulations use much of the same information
assembled for the recent UCERF3 estimates of earthquake probabilities (Field et al., 2013). These include the
California fault system model, fault slip rates, and rake angles. For this purpose we also tune the simulations to the
UCERF3 mean recurrence intervals for earthquakes at paleoseismic sites (Field et al., 2013, Appendix H). To develop
robust statistics and to sample the range of earthquake ruptures that can occur in the interacting California fault system,
we are currently setting up to run simulations that span 106 years (108 earthquakes, magnitude ~M3.5-8.0).
The recurrence intervals for large earthquakes in previous RSQSim California simulations exhibited narrower
distributions than the UCERF3-TI model (Field et al., 2015). It was expected that adding deep creeping extensions
would increase the coefficients of variation (COV) of these recurrence intervals. While the deep creeping extensions
did slightly increase the COV’s (~10%), they are still on the low end of expected values from the UCERF3-TI model.
WELCOME
2016 SCEC Annual Meeting page 73
The planned tests and suite of catalogs includes simulations with heterogeneous physical parameters, which will likely
further improve the COV’s, therefore we
do not anticipate any change in the project
timeline.
For probabilistic analysis, the fault system
is divided into sections that consist of
multiple fault boundary elements (e.g. a
section might be 10km along-strike and
extend from the surface to the base of
seismogenic zone). Each section has an
earthquake catalog from which the
density distribution for earthquake
recurrence times is constructed. Using the
standard approach with the density
distributions for recurrence times, it is a
straight-forward step to compute the time-
dependent probabilities CPi of some
earthquake in a specified magnitude
range, conditional on the time of the last
earthquake for every section i.
Potentially damaging earthquakes
typically span many sections. The relation of the section probabilities CPi to the individual event probabilities probj is
were i is an index over all fault sections, j is an index over all earthquake events in the magnitude range of the analysis,
and ij is the event participation matrix that takes values of 1 for section numbers i that participate in earthquake j and
otherwise has value =0. Generally i ≠ j and the preferred values of probj minimize the misfit error in equation (1).
Currently we are evaluating the relative merits of different optimization methods to solve for probj. Optimization using
the Bounded-Variable Least-Squares algorithm appears to be significantly more efficient than simulated annealing and
yields identical results. Figure 15.1 illustrates an example of solutions for CPi and probj using results from a short
duration trial simulation with a limited set of northern California fault set. The probabilities give the chance of an
earthquake M ≥ 6.7 for a randomly selected interval in the simulation.
15.2. RSQSim Merge to NCSA Blue Waters and Parameter Space Exploration
During its development and prior to its integration into CISM, RSQSim was initially set-up to run on TACC’s Stampede.
The code was ported to NCSA’s Blue Waters during the second quarter of 2016 through SCEC-supported NSF PRAC
allocations. Several test simulations were performed on Blue Waters. These tests verified the performance of the
earthquake simulator RSQSim with the new California fault model on the Blue Waters system, and established
estimates of the runtimes of future test simulations as well as the final, production simulations. Cross validation of
UCERF3 and RSQSim was performed within this project to determine the effect of the uncertainties in physical
parameters observed in the field and measured in the lab, on the uncertainties in probabilistic forecasting. Several
simulations were performed to investigate the effects on catalog statistics and simulation runtimes of varying the
physical parameters (i.e. rate- and state-friction parameters a, b, and Dc, the initial shear and normal stresses, and the
earthquake slip speed) used in the simulations and determine the range of values that will be set for each parameter
in the suite of catalogs that will be generated. A primary list of the planned suite of simulated catalogs was produced
from these results. As a group, we have formulated a list of metrics for use in validation of each of the catalogs. A new
California fault model has also been tested. This new, more realistic model has additional down-dip, rate-strengthening
fault extensions that are tuned to slip slowly, in response to stresses transferred by slip of the rate-weakening faults
above.
Furthermore, a new, more comprehensive method for tuning the catalogs to match observed recurrence intervals has
been designed and is currently being tested and further developed. This method uses the recurrence intervals for each
WELCOME
2016 SCEC Annual Meeting page 74
fault segment from the UCERF3 model to calibrate the normal stresses in the entire fault model in order to change the
recurrence intervals on individual fault segments in the simulated catalogs. This is an improvement on the original
tuning method which only considered the recurrence intervals for the 32 paleoseismic sites compiled for the UCERF3-
TI model, which resulted in normal stress modifications to only a small percentage of the fault elements.
15.3. Integration of RSQSim Catalogs into OpenSHA
Progress has been made integrating results from RSQSim into OpenSHA, an open-source Java-based platform for
conducting seismic hazard analysis (SHA) (Field et al., 2003). OpenSHA is the computational platform for the UCERF3
models, and was used to calculate the California component of the 2014 National Seismic Hazard Maps produced by
the USGS. RSQSim fault input files and seismicity catalog files can now be read into OpenSHA for analysis and hazard
computation. This allows for direct comparisons with best available ERF models (such as UCERF3), and we have
begun submitting RSQSim results through the same series of comprehensive tests that were used as acceptance
criteria for the UCERF3 models. This will help with the process of tuning RSQSim catalogs and asses their performance
against best available existing models. RSQSim catalogs will also provide a physics-based comparison for aspects of
the UCERF3 model which are poorly constrained by data, such as the frequency of multi-fault ruptures.
This integration also allows us to compute fully probabilistic seismic hazard curves and disaggregation through pairing
RSQSim models with the best available Ground Motion Prediction Equations (GMPEs), which estimate ground motions
for scenario earthquakes. For the Western U.S. these GMPEs were developed by the Pacific Earthquake Engineering
Research center (PEER) and are know as the NGA-West2 GMPEs (Bozorgnia et al. 2014). The NGA-West2 GMPEs
implemented in OpenSHA were thoroughly verified by a USGS/SCEC team as part of the USGS NSHMP development.
OpenSHA is also a key part of the CyberShake physics-based SHA platform (Graves et al., 2010), and integration of
RSQSim catalogs in OpenSHA is an important step toward the first end-to-end physics-based SHA calculations.
15.4. Development of Visualization Tools for UCERF3 and RSQSim Earthquake Catalog
New visualization tools have been
developed allowing scientists to
interrogate seismic sequences in the
SCEC 3-D visualization platform,
SCEC-VDO. Users can now load-in the
high-resolution triangular fault model
used as input to RSQSim, and create
rich animations of RSQSim catalogs. A
screenshot of a RSQSim rupture
visualized in the SCEC-VDO software is
shown in Figure 15.2. Users can also
render movies of full seismic sequences
of various timescales, from days to
thousands of years. An example of one
such animation of 10 thousand years of
seismicity can be downloaded here:
http://hypocenter.usc.edu/research/Keck_CISM/2016_07_12-rsqsim-10kyrs.mov.
15.5. Use of RSQSim and High-Performance Computing in Undergraduate Students Training
In addition to working towards the science goals of the project, several CISM participants were mentors for the 2016
SCEC Undergraduate Studies in Earthquake Information Technology (UseIT) program. The grand challenge for the
interns this summer was based on the CISM project and involves developing earthquake forecasts and hazard maps
for California from long-term simulated earthquake catalogs from RSQSim. Dr Gilchrist and Mr. Milner worked closely
with the High Performance Computing (HPC) team, a group of 5 undergraduates, who succeeded in running several
large RSQSim simulations on Blue Waters. These students learned to use and understand RSQSim, setup and submit
jobs to the Blue Waters supercomputer, process and evaluate the catalogs using R (a powerful open-source
programming language for data science), and identify sequences of large, particularly damaging earthquakes. Long
simulated catalogs (several hundred-thousand years) were passed to the Probabilistic Forecasting, Virtual Display of
Objects (VDO), and Hazard and Risk teams, who created animations of seismic activity, shake maps, and earthquake
WELCOME
2016 SCEC Annual Meeting page 75
forecasts for the San Andreas Fault system. These exercises provided them with real experience in the process of
large-scale scientific projects, as well as an understanding of earthquake hazard, earthquake physics, and the
applications and capabilities of high performance computing.
15.6. Select Publications
● Dieterich, J.H. (2015). "Potential Applications of RSQSim for the CISM Project", CISM Kick-off Meeting, SCEC
Annual Meeting, Palm Springs, CA, September 12-16, 2015.Gilchrist, J.J., Dieterich, J.H. and K.B. Richards-
Dinger (2016). "Update on Earthquake Forecasting with RSQSim" CISM Workshop, Los Angeles, CA, March
9 2016.
● Gilchrist, J.J., Dieterich, J.H., and K.B. Richards-Dinger (2015) "RSQSim Status Update and Potential
Contributions to OEF", Operational Earthquake Forecast Powell Meeting, Fort Collins, CO, October 19-22,
2015.
● Jordan, T.H. (2015). "Introduction and overview of CISM", FESD Workshop, U.C. Riverside, CA, April 28,
2016.
● Milner, K.R. and T.H. Jordan (2016). "Supercycles and Synchronization Signatures in Synthetic Seismic
Sequences", Seismological Society of America Annual Meeting, Reno, NV. April 20-23, 2016.
● Milner, K. R., Sanskriti, F., Yu, J., Callaghan, S., Maechling, P. J., & Jordan, T. H. (2016, 08). SCEC-VDO
Reborn: A New 3-Dimensional Visualization and Movie Making Software for Earth Science Data. SCEC
Annual Meeting, Palm Springs, CA, September 10-14, 2016.
● Pekurovsky, D., Chourasia, A., Richards-Dinger, K. B., Shaw, B. E., Dieterich, J. H., & Cui, Y. (2016, 08).
Performance enhancements and visualization for RSQSim earthquake simulator.
● Richards-Dinger, K.B. (2015). "RSQSim Code Development", CISM Kick-off Meeting, SCEC Annual Meeting,
Palm Springs, CA, September 12-16, 2015.
WELCOME
2016 SCEC Annual Meeting page 76
16. Central California Seismic Project (CCSP)
The largest uncertainties in the estimation of the catastrophic risks to California utilities come from the seismic hazard
uncertainties at low exceedance probabilities. Recent analyses indicate that these are dominated by the uncertainties
in path effects; i.e., in the prediction of strong ground motions at a fixed surface site from specified seismic sources.
SCEC has joined the Pacific Gas & Electric Company (PG&E) in developing a long-term research program aimed at
reducing the uncertainties in seismic hazard estimation with a particular emphasis of reducing the uncertainty in path
effects.A pilot project focused on the central coast of California was initiated in 2015 and continued through 2016. The
goal of the CCSP is to assess the effectiveness of physics-based ground motion modeling in reducing path-effect
uncertainties.
This past year’s accomplishments include progress on the following activities:
● Shaw and Plesch developed a Unified Structural Representation for Central California including the San
Joaquin Basin (see Section 5.2). They also improved the Santa Maria basin from the southern California USR
using industry seismic reflection and well data.
● Gill led the integration and verification of iteration six of the Central California full 3D tomography model with
UCVM
● Gill et al. developed a 1D model for Central California to be used in CyberShake
● Nakata and Beroza initiated work on the stochastic characterization of crustal structure for high-frequency
ground motion prediction using dense-array observations
● Boue’, Nakata, Denolle, and Beroza develop 9-component tensor Green’s functions to be used for input to
full-waveform tomography for the SCEC CVM.
● Archuleta and Crempien made progress on their high-frequency path and source parameters determined from
recorded ground motion in Central California
● Galvez, Bayless, and Somerville modeled backthrust faulting on the Ventura Fault System using dynamic and
kinematic modeling and the San Simeon Earthquake
● Steidl and Okaya finalized a long term deployment plan and started deployment of five new ground motion
stations
● Definition and initiation of first two CyberShake Central California calculations on NCSA Blue Waters and
OLCF Titan (Callaghan, Maechling, Jordan, Gill, Goulet)
16.1 Integration and verification of iteration six of the Central California full 3D tomography model
into UCVM, Development of a new 1D velocity model
The latest iteration of the full 3D
tomographic (F3DT) inversion of Central
California was integrated into UCVM.
David Gill (SCEC) also implemented a new
suite of tools in UCVM to allow for model
verification. New plotting tools include a
revamped set of plan views made available
at various depths and a cross-section
generation module. The visualization
allows for a quick inspection of the model
while output from computations
systematically test the properties of the
model. Figure 16.1 shows the shear wave
velocity cross-section of the model from
the Coast to the Sierras, highlighting the
deep sedimentary basin of the Great
Valley. The model was used to generate a
mesh for CyberShake computations. An
additional 1D velocity model was
generated by averaging the model properties across the domain. The 1D model is also being used for CyberShake
computations and will serve as the baseline for comparison against improved 3D models.
WELCOME
2016 SCEC Annual Meeting page 77
16.2. Dynamic and Kinematic Modeling of
Backthrust Faulting on the Ventura Fault
System: Validation using the San Simeon
Earthquake
Using 2D dynamic ruptures simulations, Galvez et
al. have generated splay ruptures where the
earthquake nucleates in the deep section of the
Ventura fault and, once the up-dip rupture reaches
the shallow section it activates and breaks the
conjugate fault (Lion backthrust). To perform stable
dynamic rupture simulations in this complex
system, the group successfully generated a stable
mesh based on the model provided by John Shaw
(Section 5.2) that includes the Ventura fault and the
Lion backthrust. The group is evaluating whether
they can extend their dynamic rupture simulations
to 3D for the Ventura fault system and still provide
results in a reasonable amount of time.
16.3. Instrumentation of Five New Ground
Motion Stations
Steidl led the instrumentation planning activities for
Central California, working closely with Okaya, to
define the sites and instrumentation packages for
50 new recording stations in the next few years
(Figure 16.2). Five new stations are being
instrumented using existing SCEC PBIC equipment
with 3 new posthole sensors budgeted for 2016.
Five more stations have been selected and
permitted for the 2017 instrumentation deployment.
For the new posthole, the group has been following
test results for Trillium Compact Posthole, Metrozet
MBB-1, and STS2 conducted at the Pasadena
Vault as part of the SCSN instrumentation program.
Steidl is working with the USGS with respect to
collaborating on permitting when possible (using
USGS permits from certain federal, state, and local
agencies) and to coordinate the
communications/telemetry efforts with respect to
data flow.
16.4 Definition and initiation of first two
CyberShake Central California
calculations (Study 16.9) on NCSA Blue
Waters and OLCF Titan
The goal of the CCSP is to improve the quantification of path effects and their uncertainty. The project is designed
around an iterative flowchart in which simple initial models are used, sophisticated models are developed, such as
those from the F3DT inversions and improvements hazard assessment can be quantified. Cybershake was selected
as the main tool for computing ground motions and for performing this assessment. Two first CyberShake runs have
been designed around CCA6 (the sixth iteration of the F3DT model) and the 1D model (Section 16.1). Figure 16.3
WELCOME
2016 SCEC Annual Meeting page 78
shows the Central California CyberShake box and the location of the 336 sites selected for the hazard map generation.
Sites are located on a 10 km grid, but optimized to correspond to recording stations, critical facilities and historic
missions for which damage from past earthquakes was quantified. CyberShake computations are performed on two
main HPC: NCSA Blue Waters and OLCF Titan. Phil Maechling and Scott Callaghan led the coordination with both
HPC centers in preparation for the two sets of computations, which is essential to ensure succesful and efficient
computations.
16.5. Select Publications
● Callaghan, S., Maechling, P. J., Goulet, C. A., Vahi, K., Graves, R. W., Olsen, K. B., Milner, K. R., Gill, D., Cui,
Y., & Jordan, T. H. (2016, 08). CyberShake Advancements: Broadband Seismograms and Central California
Sites. SCEC Annual Meeting, Palm Springs, CA, September 10-14, 2016.
● Gill, D., Maechling, P. J., Jordan, T. H., Taborda, R., Callaghan, S., & Small, P. (2013, 7). SCEC Unified
Community Velocity Model: Mesh Generation and Visualization. Oral Presentation at CIG/QUEST/IRIS Joint
Workshop on Seismic Imaging of Structure and Source.
● Gill, D., Small, P., Maechling, P. Jordan, T., Shaw, J., Pleasch, A., Lee, E.-J.,Chen, P., Goulet, C., Taborda,
R., Olsen, K., Callaghan, S. (2016) UCVM: From Supercomputers to Laptops, Querying and Visualizing 3D
Seismic Velocity Models, American Geophysical Union Fall Meeting, San Francisco, CA, December 12-16,
2016.
● Shaw, J. H., A. Plesch, C. Tape, M. P. Suess, T. H. Jordan, G. Ely, E. Hauksson, J. Tromp, T. Tanimoto, R.
Graves, K. Olsen, C. Nicholson, P. J. Maechling, C. Rivero, P. Lovely, C. M. Brankman, J. Munster, (2015).
Unified Structural Representation of the southern California crust and upper mantle, Earth and Planetary
Science Letters, 415, 1–15.
WELCOME
2016 SCEC Annual Meeting page 79
Draft 2017 Science Plan
SCEC Science Planning Committee, September 2016
1 New This Year
The 2017 SCEC Science Plan (aka RFP) reflects research priorities articulated in the SCEC5 proposal. Substantial
changes have been made to the RFP, so we strongly encourage researchers to read carefully the RFP in its entirety.
● The Science Planning Committee (PC) is reconfigured for SCEC5. Some working groups remain, some have
checked out, and some have evolved to reflect a changed emphasis. Specifically, the four disciplinary
committees (Seismology, Geodesy, Geology, and Computational Science), as well as the interdisciplinary
focus groups (FARM, SDOT, and EERI) are the same. GMP and SoSAFE change slightly to the more general
Ground Motions and San Andreas Fault System, respectively. The USR focus group will broaden to include
all community models under the new CXM group. Finally, Special Projects will no longer be represented
individually on the PC, but rather in aggregate by the Executive Director for Special Projects (Christine Goulet)
and the SCEC IT Architect (Phil Maechling).
● All Technical Activity Groups (TAGs) will sunset at the end of SCEC4. Established TAGs, and proposed new
TAGs, will have to be re-initiated in SCEC5 through submission of a successful proposal to the Planning
Committee.
● The San Gorgonio Pass and Ventura Special Fault Study Areas are wrapping up at the end of SCEC4. This
does not mean that we will no longer fund proposals that concern research in these areas. The evolved
approach for focused multidisciplinary research within SCEC will be under the new Earthquake Gates
Initiative.
● Investigators on proposals that anticipate use of SCEC computational resources and/or help from SCEC
software developers should consult with SCEC IT leadership for budget time support estimates and
coordination planning.
● Investigators interested in undergraduate summer interns should include an "intern project" description in their
proposal. The undergraduate intern will be recruited by the SCEC/CEO Program staff. Selected intern projects
will be awarded as supplemental funds on the proposal award. Funds will be disbursed and managed at USC
to use for a summer stipend and travel support to the SCEC annual meeting for the selected undergraduate
student. The number of intern projects awarded each year will depend on available funding in the SCEC
annual budget and the available applicant interest pool.
2 Overview
The Southern California Earthquake Center (SCEC) was founded as a Science & Technology Center on February 1,
1991, with joint funding by the National Science Foundation (NSF) and the U. S. Geological Survey (USGS). Since
2002, SCEC has been sustained as a stand-alone center under cooperative agreements with both agencies in three
consecutive, five-year phases (SCEC2-SCEC4). The Center has been extended for another 5-year period, effective 1
Feb 2017 to 31 Jan 2022 (SCEC5). SCEC coordinates fundamental research on earthquake processes using Southern
California as its main natural laboratory. Currently, over 1000 earthquake professionals are participating in SCEC
projects. This research program is investigator-driven and supports core research and education in seismology, tectonic
geodesy, earthquake geology, and computational science. The SCEC community advances earthquake system
science by gathering information from seismic and geodetic sensors, geologic field observations, and laboratory
experiments; synthesizing knowledge of earthquake phenomena through system-level, physics-based modeling; and
communicating understanding of seismic hazards to reduce earthquake risk and promote community resilience.
2.1 The SCEC5 Research Vision
Earthquakes are emergent phenomena of active fault systems, confoundingly simple in their gross statistical features
but amazingly complex as individual events. SCEC’s long-range science vision is to develop dynamical models of
earthquake processes that are comprehensive, integrative, verified, predictive, and validated against observations. The
WELCOME
2016 SCEC Annual Meeting page 80
science goal of the SCEC5 core program is to provide new concepts that can improve the predictability of the
earthquake system models, new data for testing the models, and a better understanding of model uncertainties.
The validation of model-based predictions against data is a key SCEC activity, because empirical testing is the most
powerful guide for assessing model uncertainties and moving models towards better representations of reality. SCEC
validation efforts tightly couple basic earthquake research to the practical needs of probabilistic seismic hazard analysis,
operational earthquake forecasting, earthquake early warning, and rapid earthquake response. Moreover, the risk-
reduction problem—which requires actions motivated by useful information—strongly couples SCEC science to
earthquake engineering. SCEC collaborations with engineering organizations are directed towards end-to-end, physics-
based modeling capabilities that span system processes from “ruptures-to-rafters.”
SCEC connects to the social sciences through its mission to convey authoritative information to stakeholders in ways
that result in lowered risk and enhanced resilience. SCEC’s vision is to engage end-users and the public at large in on-
going, community-centric conversations about how to manage particular risks by taking specific actions. The SCEC
Communication, Education, and Outreach (CEO) program seeks to promote this dialog on many levels, through many
different channels, and inform the conversations with authoritative earthquake information. Towards this goal, the
SCEC5 CEO program continues to build networks of organizational partners that can act in concert to prepare millions
of people of all ages and socioeconomic levels for inevitable earthquake disasters.
2.2 The SCEC5 Science Plan
The SCEC5 Science Plan was developed by the non-USGS members of the SCEC Planning Committee and Board of
Directors with extensive input from issue-oriented “tiger teams” and the community at large. The strategic framework
for the SCEC5 Science Plan is cast in the form of five basic questions of earthquake science: (1) How are faults loaded
on different temporal and spatial scales? (2) What is the role of off-fault inelastic deformation on strain accumulation,
dynamic rupture, and radiated seismic energy? (3) How do the evolving structure, composition and physical properties
of fault zones and surrounding rock affect shear resistance to seismic and aseismic slip? (4) How do strong ground
motions depend on the complexities and nonlinearities of dynamic earthquake systems? (5) In what ways can system-
specific studies enhance the general understanding of earthquake predictability? These questions cover the key issues
driving earthquake research in California, and they provide a basis for gauging the intellectual merit of proposed SCEC5
research activities.
Research priorities have been developed to address these five basic questions. Tied to the priorities are fourteen
science topics distributed across four main thematic areas.
2.2.1 Basic Questions of Earthquake Science
Q1. How are faults loaded across temporal and spatial scales?
Problem Statement: Fault systems are externally loaded, primarily by the relatively steady forces of plate tectonics, but
also by mass transfers at the surface due to long-term interactions of the solid Earth with its fluid envelopes (climate
forcing) and by short-term gravitational interactions (tidal forcing). Much is yet to be learned about the stress states
acting on active faults and how these stress states evolve through external loading and the internal transfer of stress
during continuous deformation and discontinuous faulting.
In SCEC4, we initiated research on a Community Stress Model (CSM) to describe our current knowledge about the
stress state of the San Andreas fault system. The ensemble of stress and stress-rate models comprised by the current
CSM is a quantitative representation of how well we have been able to answer Q1. Empirical models have been
developed for stress orientations in the upper crust based on abundant focal mechanisms and more limited in-situ data,
as well as 3D dynamic models of stress; e.g., from finite-element simulations of long-term tectonics, including nonlinear
laboratory rheologies. A new approach builds 3D stress models as sums of analytic solutions that satisfy momentum
conservation everywhere, while approximating the previous stress-direction and stress-amplitude models in a least-
squares sense. Though we are encouraged by our recent progress, understanding stress is a long-term proposition.
Research Priorities:
P1.a. Refine the geologic slip rates on faults in Southern California, including offshore faults, and optimally combine
the geologic data with geodetic measurements to constrain fault-based deformation models, accounting for
observational and modeling uncertainties. (Geology, Geodesy, SDOT, SAFS)
WELCOME
2016 SCEC Annual Meeting page 81
P1.b. Determine the spatial scales at which tectonic block models (compared to continuum models) provide
descriptions of fault-system deformation that are useful for earthquake forecasting. (SDOT, Geodesy, EFP, CXM)
P1.c. Constrain how absolute stress and stressing rate vary laterally and with depth on faults, quantifying model
sensitivity, e.g. to rheology, with inverse approaches. (SDOT, CXM, Geology)
P1.d. Quantify stress heterogeneity on faults at different spatial scales, correlate the stress concentrations with
asperities and geometric complexities, and model their influence on rupture initiation, propagation, and arrest.
(Seismology, SDOT, FARM, Geology)
P1.e. Evaluate how the stress transfer among fault segments depends on time, at which levels it can be approximated
by quasi-static and dynamic elastic mechanisms, and to what degree inelastic processes contribute to stress evolution.
(SDOT, Geodesy, Seismology, FARM, CS)
Q2. What is the role of off-fault inelastic deformation on strain accumulation, dynamic rupture, and radiated
seismic energy?
Problem Statement: In the brittle upper crust, observations of low-velocity zones associated with active seismogenic
faults, together with time-dependent evolution of seismic velocities following stress perturbations suggest intrinsic
relationships between damage, healing, and effective elastic moduli of rocks in a fault zone. Such relationships are
only poorly understood, but they can elucidate the development and evolution of fault zones in space and time, as well
as the interplay between damage accumulated over multiple earthquake cycles and rupture dynamics. Current dynamic
rupture models show that the assumption of elastic deformation of the host rocks is often violated; e.g., in regions of
high stress concentration near the propagating rupture front, particularly when stress is further concentrated by
geometrically complex fault surfaces. This raises important questions about the effect of nonlinearity and damage on
the nucleation, propagation, and arrest of rupture. Neglecting inelastic response may systematically bias inversions of
seismic and geodetic data for slip distribution and rupture geometry, affect measurements of coseismic slip at the
surface, and inferences of long-term slip rates from the geologic record.
The SCEC community is at the forefront of research on inelastic material response associated with earthquake faulting
and its effects on dynamic rupture propagation and seismic ground motion. The SCEC focus on extreme ground motion
for the Yucca Mountain Project drew attention to the physical limits that realistic, inelastic material response places on
strong shaking. Recent simulations of earthquakes in the Los Angeles region have demonstrated how yielding near the
fault and in sedimentary basins substantially reduces predicted ground motions relative to purely elastic simulations.
Accounting for inelasticity brings the model predictions more in line with empirical constraints on strong shaking.
Research Priorities:
P2.a. Determining how much off fault plasticity contributes to geodetic estimates of strain accumulation and what
fraction of seismic-moment accumulation is relaxed by aseismic processes. (FARM, Geodesy, CS)
P2.b. Explore approaches to represent the effects of non-linearity that would allow the continued use of linear wave
propagation as an effective approximation. (GM, CS, Seismology)
P2.c Constrain the form of fault-zone and distributed non-linearity, as well as the factors, such as cohesion and pore
fluid pressure, that are likely to influence it. (FARM, CS, GM, Seismology)
P2.d. Understand how inelastic strain associated with fault roughness and discontinuities influences rupture
propagation, seismic radiation, and scaling of earthquake source parameters. (CS, FARM, Seismology)
P2.e Describe how fault complexity and inelastic deformation interact to determine the probability of rupture propagation
through structural complexities, and determine how model-based hypotheses about these interactions can be tested
by the observations of accumulated slip and paleoseismic chronologies. (EFP, FARM, CS, Geology, SAFS)
Q3. How do the evolving structure, composition and physical properties of fault zones and surrounding rock
affect shear resistance to seismic and aseismic slip?
Problem Statement: Fault systems show complexities that range from the macroscales of plate tectonics to the
microscales of highly damaged rocks that are fluid-filled and chemically reactive. Many questions about the evolving
dynamics of these complex systems remain unanswered. The inferred values of heat outflow from mature faults, such
as the San Andreas, Taiwan’s Chelungpu Fault, and the Japan Trench megathrust, imply that shear stress acting during
WELCOME
2016 SCEC Annual Meeting page 82
sliding is an order of magnitude lower than estimates from Byerlee’s law and typical static friction measurements—an
inconsistency famously known as the “heat-flow paradox.” Low values for shear stress acting on major faults are also
supported by the steep angles between the principal stress direction and fault trace, slip-vector rake rotations during
faulting, and significant rotations of principal stresses after large earthquakes. In addition, multi-fault earthquake
simulations show that observed propagation onto unfavorably oriented structures appears to be more likely to occur if
the faults are subject to low tectonic stress.
These and other observations motivate the continued investigation of the structure, composition, and physical
properties of fault zones that host earthquake sources. One important question is which faults are susceptible to
coseismic weakening mechanisms, such as flash heating, thermal pressurization of pore fluids, partial or full melting of
the shearing zone, silica-gel formation, and thermal decomposition of sheared materials into friction-reducing
byproducts. Co-seismic weakening may lead to large unexpected slip in creeping fault regions, including deeper fault
extensions below the seismogenic layer, a phenomenon compatible with a range of observations. Fluids play a key
role in several of the weakening processes, potentially dominating co-seismic resistance to slip. In fact, fluids can lead
to extreme localization of the shearing layer, promoting co-seismic weakening. Fluids can also provide a stabilizing
factor, for example due to inelastic shear-induced dilatancy of the pore space, and the resulting reduction of pore
pressure and hence increase of the effective normal stress.
Research Priorities:
P3.a. Refine the geometry of active faults across the full range of seismogenic depths, including structures that link and
transfer deformation between faults. (CXM, Seismology, Geodesy, Geology, SAFS)
P3.b. Constrain the active geometry and rheology of the ductile roots of fault zones. (CXM, SDOT, FARM, Geology)
P3.c. Assess how shear resistance and energy dissipation depend on the maturity of the fault system, and how these
are expressed geologically. (FARM, SDOT, Geology)
P3.d. Determine how damage zones, crack healing and cementation, fault zone mineralogy, and off-fault plasticity
govern strain localization, the stability of slip (creeping vs. locked), interseismic strength recovery, and rupture
propagation. (FARM, Geology, CS)
P3.e Constrain the extent of permanent, off-fault deformation, and its contribution to geologic and geodetic fault slip-
rate estimates. (Geology, SAFS, Geodesy, EFP)
P3.f. Study the mechanical and chemical effects of fluid flows, both natural and anthropogenic, on faulting and
earthquake occurrence, and how they vary throughout the earthquake cycle. (FARM, Geology, EFP)
P3.g. Assess the importance of the mechanical properties of the near-surface in the commensurability of geodetic and
seismological images of fault slip at depth with fault offset expressed at the surface. (Seismology, Geodesy, Geology,
GM, FARM)
Q4. How do strong ground motions depend on the complexities and nonlinearities of dynamic earthquake
systems?
Problem Statement: Physics-based predictions of strong ground motions are “the proof of the pudding”; comparing
them with data is essential to testing our understanding of source and wave dynamics, and they connect the basic
science of earthquakes to the practical applications of seismic hazard analysis. Ground-motion simulations have
become useful in performance-based engineering and nonlinear building response analysis, operational earthquake
forecasting, and earthquake early warning. The use of validated numerical simulations can yield predictions adapted
to local geologic conditions such as sedimentary basins, structural boundaries, and steep topography, and they provide
meaningful ground-motion estimates for conditions poorly represented in the empirical database.
An appropriate baseline for measuring future progress in ground-motion modeling is the recent CyberShake 15.4 study,
which produced hazard curves for the Los Angeles region from a stochastically complete set of UCERF2 ruptures using
the CVM-S4.26 crustal structure. The resulting hazard model has several notable limitations: (i) the sources were
prescribed by a pseudo-dynamic (kinematic), rather than fully dynamic, rupture model; (ii) the wavefield calculations
were computed to an upper cutoff frequency of 1 Hz, compared to engineering needs that can exceed 10 Hz; and (iii)
the principle of seismic reciprocity was used to compute the requisite ensemble of seismograms. To preserve
reciprocity, which strictly applies only to perfectly elastic media, near-fault inelasticity would have to be built into the
WELCOME
2016 SCEC Annual Meeting page 83
rupture model a priori as a source effect, whereas near-surface inelasticity would have to be incorporated a posteriori
as a site effect.
We seek to replace the classical treatment of source, path, and site effects as decoupled processes by a new paradigm
in which the surface ground motions are modeled as the nonlinear response of a self-excited dynamical system with a
rheology that properly represents the most salient aspects of inelastic behavior. As the CyberShake example indicates,
this will be a major challenge for SCEC5. Our plan will be guided by four priorities that recognize the practical potential
of this paradigm shift.
Research Priorities:
P4.a. Determine the relative roles of fault geometry, heterogeneous frictional resistance, wavefield scattering, intrinsic
attenuation, and near-surface nonlinearities in controlling ground motions. (GM, Seismology, EEII)
P4.b. Construct methods for validating ground-motion predictions that account for the paucity of recordings in the near-
field, where the motions are strong and inelastic effects may be large. (GM, EEII)
P4.c. Develop ground-motion simulations for anticipated large events that are suitable for probabilistic seismic hazard
and risk analysis. (CS, GM, EEII)
P4.d. Communicate improvements in physics-based seismic hazard analysis to the earthquake engineers, emergency
responders, and general public. (EEII, GM)
Q5. In what ways can system-specific studies enhance the general understanding of earthquake predictability?
Problem Statement: Earthquake prediction is one of the great unsolved problems of physical science. We distinguish
intrinsic predictability (the degree to which a future earthquake behavior is encoded in the precursory behavior of an
active fault system) from a specific prediction (a testable hypothesis, usually stated in probabilistic terms, of the location,
time, and magnitude of an earthquake). A key objective of the SCEC5 core program is to improve our understanding
of earthquake predictability as the basis for advancing useful forecasting models. We propose to take a broad view of
the earthquake predictability problem. For example, many interesting problems of conditional predictability can be
posed as physics questions in a system-specific context. What will be the shaking intensity in the Los Angeles basin
from a magnitude 7.8 earthquake on the southern San Andreas Fault? By how much will the strong shaking be amplified
by the coupling of source directivity to basin effects? Will deep injection of waste fluids cause felt earthquakes near a
newly drilled well in the San Joaquin Valley? How intense will the shaking be during the next minute of an ongoing
earthquake in Los Angeles?
Earthquake system science offers a “brick-by-brick” approach to improving our understanding earthquake predictability.
In SCEC5, we propose to build system-specific models of rupture recurrence, stress evolution, and triggering within a
probabilistic framework that can assimilate a wide variety of geologic, geodetic, and seismic observations. Five research
priorities will guide this plan.
Research Priorities:
P5.a. Develop earthquake simulators that encode the current understanding of earthquake predictability. (EFP, CS,
FARM)
P5.b. Place useful geologic bounds on the character and frequency of multi-segment and multi-fault ruptures of extreme
magnitude. (SAFS, Geology, EFP)
P5.c. Assess the limitations of long-term earthquake rupture forecasts by combining patterns of earthquake occurrence
and strain accumulation with neotectonic and paleoseismic observations of the last millennium. (EFP, SAFS, Geology)
P5.d. Test the hypothesis that “seismic supercycles’’ seen in earthquake simulators actually exist in nature and explore
the implications for earthquake predictability. (EFP, SAFS, Geology)
P5.e. Exploit anthropogenic (induced) seismicity as experiments in earthquake predictability. (FARM, EFP)
2.2.2 SCEC5 Thematic Areas and Topical Elements
The basic science questions reflect the core issues currently driving earthquake research. SCEC5 will address these
questions through an interdisciplinary program comprising 14 topics in four main thematic areas. While these are by no
WELCOME
2016 SCEC Annual Meeting page 84
means the only research activities to be undertaken in SCEC5, they constitute a cogent plan for making progress on
the core scientific issues.
1. Modeling the fault system: We seek to know more about the geometry of the San Andreas system as a complex
network of faults, how stresses acting within this network drive the deformation that leads to fault rupture, and how this
system evolves on time scales ranging from milliseconds to millions of years.
● Stress and Deformation Over Time. We will build alternative models of the stress state and its evolution during
seismic cycles, compare the models with observations, and assess their epistemic uncertainties, particularly
in the representation of fault-system rheology and tectonic forcing.
● Earthquake Gates. Earthquake gates are regions of fault complexity conjectured to inhibit propagating
ruptures, owing to dynamic conditions set up by proximal fault geometry, distributed deformation, and
earthquake history. We will test the hypothesis that earthquake gates control the probability of large, multi-
segment and multi-fault ruptures.
● Community Models. We will enhance the accessibility of the SCEC Community Models, including the model
uncertainties. Community thermal and rheological models will be developed.
● Data Intensive Computing. We will develop methods for signal detection and identification that scale efficiently
with data size, which we will apply to key problems of Earth structure and nanoseismic activity.
2. Understanding earthquake processes: Many important achievements in understanding fault-system stresses, fault
ruptures, and seismic waves have been based on the elastic approximation, but new problems motivate us to move
beyond elasticity in the investigation of earthquake processes.
● Beyond Elasticity. We will test hypotheses about inelastic fault-system behavior against geologic, geodetic,
and seismic data, refine them through dynamic modeling across a wide range of spatiotemporal scales, and
assess their implications for seismic hazard analysis.
● Modeling Earthquake Source Processes. We will combine co-seismic dynamic rupture models with inter-
seismic earthquake simulators to achieve a multi-cycle simulation capability that can account for slip history,
inertial effects, fault-zone complexity, realistic fault geometry, and realistic loading.
● Ground Motion Simulation. We will validate ground-motion simulations, improve their accuracy by
incorporating nonlinear rock and soil response, and integrate dynamic rupture models with wave-scattering
and attenuation models. We seek simulation capabilities that span the main engineering band, 0.1-10 Hz.
● Induced Seismicity. We will develop detection methods for low magnitude earthquakes, participate in the
building of hydrological models for special study sites, and develop and test mechanistic and empirical models
of anthropogenic earthquakes within Southern California.
3. Characterizing seismic hazards: We seek to characterize seismic hazards across a wide spectrum of anticipation
and response times, with emphasis on the proper assessment of model uncertainties and the use of physics-based
methods to lower those uncertainties.
● Probabilistic Seismic Hazard Analysis. We will attempt to reduce the uncertainty in PSHA through physics-
based earthquake rupture forecasts and ground-motion models. A special focus will be on reducing the
epistemic uncertainty in shaking intensities due to 3D along-path structure.
● Operational Earthquake Forecasting. We will conduct fundamental research on earthquake predictability,
develop physics-based forecasting models in the new Collaboratory for Interseismic Simulation and Modeling,
and coordinate the Working Group on California Earthquake Probabilities.
● Earthquake Early Warning. We will develop methods to infer rupture parameters from time-limited data,
ground-motion predictions that account for directivity, basin, and other 3D effects, and better long-term and
short-term earthquake rupture forecasts for conditioning of early-warning algorithms.
● Post-Earthquake Rapid Response. We will improve the rapid scientific response to strong earthquakes in
Southern California through the development of new methods for mobilizing and coordinating the core
geoscience disciplines in the gathering and preservation of perishable earthquake data.
4. Reducing seismic risk: Through partnerships coordinated by SCEC’s Earthquake Engineering Implementation
Interface, we will conduct research useful in motivating societal actions to reduce earthquake risk. Two topics
investigated by these engineering partnerships will be:
● Risk to Distributed Infrastructure. We will work with engineers and stakeholders to apply measures of
distributed infrastructure impacts in assessing correlated damage from physics-based ground-motion
simulations. An initial project will develop earthquake scenarios for the Los Angeles water supply.
WELCOME
2016 SCEC Annual Meeting page 85
● Earthquake Physics of the Geotechnical Layer. In collaboration with geotechnical engineers, we will advance
the understanding of site effects and soil-structure interactions by incorporating nonlinear rheological models
of near-surface rock and soil layers into full-physics earthquake simulations.
2.3 Management of the SCEC Research Program
The SCEC Science Planning Committee (PC) is responsible for developing the SCEC Annual Science Plan, which
describes the Center’s research interests and priorities, and the SCEC Annual Collaboration Plan, which details how
resources will be allocated to projects. The PC is chaired by the SCEC Co-Director (Greg Beroza), who is assisted by
a PC Vice-Chair (Judi Chester). The PC comprises the leaders of the SCEC science working groups—disciplinary
committees, focus groups, and special project groups—who, together with the working group co-leaders, guide SCEC’s
research program. The Executive Director for Special Projects (Christine Goulet) manages science activities in the
externally funded SCEC Special Projects and coordinates these activities with PC and the SCEC IT Architect (Phil
Maechling), who has oversight of SCEC Information Technology, including the software standards for data structures
and model interfaces. The Center Director (Tom Jordan) and the Associate Directors for Administration (John
McRaney) and CEO (Mark Benthien) will serve ex officio members. The PC is responsible for formulating the Center’s
science plan, conducting proposal reviews, and recommending projects to the Board of Directors for SCEC support.
Its members play key roles in implementing the SCEC5 science plans. With the beginning of SCEC5, we have
restructured the PC – reducing it slightly in size, evolving some groups, and changing the membership significantly in
an effort to both revitalize the PC, and to ensure that it is aligned with the goals of SCEC5.
3 Annual SCEC Science Plan
The SCEC Science Plan solicits proposals from individuals and groups to participate in the SCEC research program
on an annual basis. Typical grants awarded under the SCEC Science Plan fall in the range of $10,000 to $35,000. This
is not intended to limit SCEC to a fixed award amount, nor to a specified number of awards, but rather to calibrate
expectations for proposals submitted to SCEC. Field research investigations outside southern California are generally
not supported.
3.1 Investigator Eligibility and Responsibilities
3.1.1 Who May Submit Proposals
Proposals can be submitted by eligible Principal Investigators (PI) from U.S. academic institutions and U.S. private
corporations.
SCEC rarely provides support to foreign institutions. Collaborative projects involving U.S. and foreign organizations will
be considered, provided funding is requested only for the U.S. portion of the collaborative effort.
Collaborative proposals with investigators from the U.S. Geological Survey are encouraged. USGS employees should
submit requests for support through USGS channels.
Any person with an overdue project report (for prior SCEC-funded awards) at the time of the proposal deadline will not
be allowed to submit a new or continuation proposal as a PI or co-PI.
3.1.2 Investigator Responsibilities
By submitting a proposal to SCEC, investigators agree to all three conditions listed below. Investigators who fail to meet
these conditions may (a) not receive funding until conditions are satisfied and/or (b) become ineligible to submit a future
proposal to SCEC.
1. Community Participation. Principal investigators will interact with other SCEC scientists on a regular basis and
contribute data, results, and/or models to the appropriate SCEC resource.
SCEC Annual Meeting. The PI will attend the annual meeting and present results of SCEC-funded research
in the poster sessions, workshops and/or working group meetings.
Data Sharing. Funded investigators are required to contribute data and results to the appropriate SCEC
resource (e.g., Southern California Earthquake Data Center, database, community model).
2. Project Reporting. Principal investigators will submit a project report by the due date listed below.
WELCOME
2016 SCEC Annual Meeting page 86
Workshop Awards. A report on results and recommendations of the workshop funded by SCEC is due no later
than 30 days following the completion of the workshop. The report will be posted on the SCEC website as
soon as possible after review by SCEC leadership.
SCEC Annual Meeting Participation Awards. A report on results and recommendations from travel funded by
SCEC is due no later than 30 days following the completion of the travel.
Research Awards. Investigators funded by SCEC must submit a project report no later than March 15 (5:00
pm PST) in the year after the funding was received. Reports should be a maximum of 5 pages (including text
and figures). Reports must include references to all SCEC publications during the past year (including papers
submitted and in review) with their SCEC contribution number (see 3 below).
3. Registration of Publications. Principal investigators will register publications resulting entirely or partially from
SCEC funding in the SCEC Publications System (www.scec.org/publications) to receive a SCEC contribution number.
Publications resulting from SCEC funding should acknowledge SCEC and include the SCEC contribution number.
3.2 Proposal Categories
3.2.1 Research Proposals
A. Data Gathering and Products. SCEC coordinates an interdisciplinary and multi-institutional study of
earthquakes in Southern California, which requires data and derived products pertinent to the region.
Proposals in this category should address the collection, archiving and distribution of data, including the
production of SCEC community models that are online, maintained, and documented resources for making
data and data products available to the scientific community.
B. Integration and Theory. SCEC supports and coordinates interpretive and theoretical investigations on
earthquake problems related to the Center’s mission. Proposals in this category should be for the integration
of data or data products from category A, or for general or theoretical studies. Proposals in categories A and
B should address one or more of the basic questions of earthquake science (see Questions), and may include
a brief description (<200 words) as to how the proposed research and/or its results might be used in a special
initiative (see Special Projects) or in education and/or outreach (see CEO).
3.2.2. Collaborative Proposals and Technical Activity Groups
A. Collaborative Proposals involving multiple investigators and/or institutions are strongly encouraged. The
lead investigator should submit only one proposal for the collaborative project. Information on all investigators
requesting SCEC funding (including budgets, current and pending support statements) must be included in
the proposal submission. Note that funding for Collaborative Proposals may be delayed or denied if any of the
investigators has overdue project report(s) for prior SCEC award.
B. Technical Activity Groups (TAGs) self-organize to develop and test critical methodologies for solving
specific problems. TAGs have formed to verify the complex computer calculations needed for wave
propagation and dynamic rupture problems, to assess the accuracy and resolving power of source inversions,
and to develop geodetic transient detectors and earthquake simulators. TAGs share a modus operandi: the
posing of well-defined “standard problems”, solution of these problems by different researchers using
alternative algorithms or codes, a common cyberspace for comparing solutions, and meetings to discuss
discrepancies and potential improvements.
3.2.3 Participant Support Proposals
A. Workshops. SCEC participants who wish to convene a workshop between February 1, 2017 and January 31,
2018 should submit a proposal for the workshop in response to this Collaboration Plan. The proposed lead
convener of the workshop must contact Tran Huynh ([email protected]) for guidance in planning the scope,
budget and scheduling of the proposed workshop before completing the proposal submission. Note that
workshops scheduled in conjunction with the SCEC Leadership Retreat (June) or SCEC Annual Meeting
(September) are limited in number and may have further constraints due to space and time availability.
B. SCEC/SURE Intern Project Supplement. SCEC coordinates a Summer Undergraduate Research
Experience (SURE) Program that enables undergraduate students to work alongside SCEC scientists on a
variety of research projects during the summer. Investigators interested in mentoring an undergraduate
student over the summer should include an "intern project" description in their proposal that aligns with the
overall proposed project plan. The undergraduate intern will be recruited by the SCEC Communication,
WELCOME
2016 SCEC Annual Meeting page 87
Education, and Outreach (CEO) Program staff. Selected intern projects will be awarded as supplemental funds
on the proposal award. The supplemental funds will provide (1) a summer stipend up to $5,000 for the selected
undergraduate student and (2) travel support for the student to participate in the all intern field trip and to
present his/her summer research at the SCEC annual meeting. The number of intern projects awarded each
year will depend on available funding in the SCEC annual budget and the available student applicant (interest)
pool. The funds will be disbursed and managed at USC. Questions about the SCEC/SURE Program should
be directed to Mark Benthien (213-740-0323; [email protected]).
C. SCEC Annual Meeting Participation. Eligible investigators may request travel funding to present research
relevant to the SCEC science plan at the SCEC Annual Meeting. To ensure appropriate representation and/or
availability of particular expertise in the SCEC community, travel awards to foreign individuals with established
collaborations will be considered. Requests for travel support to the SCEC Annual Meeting must be cost-
shared by the proposer’s institution. Cost-sharing must be described in the proposal.
3.3 Guidelines for Proposal Submission
3.3.1 Proposal Due Date
The deadline date is Tuesday, November 15, 2016 (5:00pm PST). Late proposals will not be accepted.
3.3.2 How to Submit Proposals
Every investigator listed on the proposal must have a registered account on www.SCEC.org, with current contact
information and profile information updated. Proposals must be submitted through the SCEC Proposal System,
accessible via www.scec.org/scienceplan.
Proposals do not need to be formally signed by institutional representatives.
3.3.3 Project Duration
The proposed project period should be 1-year duration (starting Feb 1, 2017 and ending Jan 31, 2018).
3.3.4 Proposal Contents
Every proposal submitted must include all of the contents listed below. Proposals must be received through the online
system by the due date, with all all required information, to be considered complete. Incomplete proposals may be
rejected and returned without comment.
1. Cover Page. The proposal cover page should include all the following information, which will be required when
submitting the proposal online:
● Project Title, Principal Investigator(s), and Institutional Affiliation(s)
● Total Amount of Request on Proposal, Amount of Request per Investigator
● Proposal Category (see Section 3)
● Three SCEC science objectives, ranked in order of priority, that the proposal addresses (e.g. P4.c, P3.d and
P2.a; see Section 2).
● Include the words “SCEC/SURE Intern Project” if requesting supplemental funding through the SCEC/SURE
Program
2. Project Plan. In 5 pages maximum (including figures), describe the proposed project and how it relates to SCEC5
objectives and priorities (see Section 2). References are excluded from the 5-page limit.
● Continuation Projects. If the proposed project is a continuation of a prior SCEC award, the project plan must
include a 1-page summary of the research results obtained from that SCEC funding. The research summary
is counted towards the 5-page limit.
● Collaborative Proposals. The project plan may include one extra page per investigator to report recent results
from previously-funded, related research.
● Workshop Proposals. Workshop proposals that include travel support for international participants must clearly
state how such participants are critical to the workshop.
● SURE Intern Project Supplement. The project must include an “intern project” description (1 page maximum)
if requesting support for a summer undergraduate intern. In the description briefly describe the project,
location(s) where research will take place, the role of the intern in the project, and any specific skills or
WELCOME
2016 SCEC Annual Meeting page 88
educational background required. The project description submitted will be used to recruit students and posted
on the SCEC Internship website. This description is not included in the 5-page limit.
3. Budget and Budget Justification. Every proposal must include a budget table and budget justification for each
institution requesting funding. The budgets should be constructed using NSF categories.
● Budget Guidance. Typical SCEC awards range from $10,000 to $35,000. This is not intended to limit SCEC
to a fixed award amount, nor to a specified number of awards, but rather to calibrate expectations for proposals
written by SCEC investigators.
● CCSP Budgets. The master agreement funding CCSP limits indirect costs to 15%. Use only this rate on
budgets for proposals responding to CCSP.
● Field Research. Field research investigations outside southern California are generally not supported.
● Salary Support. An investigator will receive no more than 1 month of summer salary support in any given year
from all combined SCEC funded awards in that year.
● SCEC IT Support. Investigators on proposals that anticipate use of SCEC computational resources and/or
help from SCEC software developers should consult with SCEC IT leadership for budget time support
estimates and coordination planning.
● Unallowable Direct Expenses. Under guidelines of the SCEC Cooperative Agreements and A-81 regulations,
secretarial support and office supplies are not allowable as direct expenses.
● For each organization requesting funding, the budget information must be entered through the online
submission system. This information should also be included in the PDF uploaded at the time of proposal
submission.
4. Current and Pending Support. Statements of current and pending support must be included for each Principal
Investigator requesting funding on the proposal, following NSF guidelines. Proposals without a current and pending
support statement may be rejected and returned without review.
● Each investigator requesting funding must enter his/her current and pending support information through the
online submission system. This information should also be included in the PDF uploaded at the time of
proposal submission.
● Workshop Proposals. Current and pending support information is not required from investigators submitting
workshop proposals.
3.4 Proposal Review Process
Structuring of the SCEC science plan begins with discussions at the SCEC Leadership Retreat in June every year, and
continues at the Annual Meeting in September. A Science Plan document is issued in October and proposals are
submitted to SCEC in November of each year.
In December, proposals are sent out for review. Every proposal is independently reviewed by the Center Director (Tom
Jordan), Center Co-Director/Science Planning Committee Chair (Greg Beroza), and Science Planning Committee Vice-
Chair (Judi Chester). Each proposal is also reviewed by the leaders of three relevant SCEC science working groups
(primary, secondary, and tertiary) – totaling five or more independent reviews for each proposal submitted. Reviewers
recuse themselves when conflicts of interest arise. Proposals are also reviewed by USGS members of the SCEC/USGS
Joint Planning Committee (JPC).
The SCEC Planning Committee (PC), chaired by the Center Co-Director, meets the following January to review and
discuss all submitted proposals. USGS members of the JPC also participate in the proposal review meeting. Based on
a combination of reviews and full panel discussion, the PC assigns a rating and recommended a funding level for each
proposal.
The Planning Committee's objective each year is to formulate a coherent science/infrastructure program consistent
with SCEC's basic mission given the Center's short-term objectives, long-term goals, and institutional composition, and
recommend a budget given the available funding.
Proposals are evaluated on the criteria below (not necessarily in order of priority):
1. Scientific merit of the proposed research
2. Competence and performance of the investigators, especially in regard to past SCEC-sponsored research
3. Priority of the proposed project for short-term SCEC objectives as stated in the annual collaboration plan
WELCOME
2016 SCEC Annual Meeting page 89
4. Promise of the proposed project for contributing to long-term SCEC goals as reflected in the SCEC science
plan
5. Commitment of the principal investigator and institution to the SCEC mission
6. Value of the proposed research relative to its cost
7. Ability to leverage the cost of the proposed research through other funding sources
8. Involvement of students and junior investigators
9. Involvement of women and underrepresented groups
10. Innovative or "risky" ideas that have a reasonable chance of leading to new insights or advances in earthquake
physics and/or seismic hazard analysis
11. The need to achieve a balanced budget while maintaining a reasonable level of scientific continuity given very
limited overall center funding
It is important to note that proposals receiving a low rating or no funding does not necessarily imply a scientifically
inferior proposal. Rather, proposals may be downgraded based on other criteria noted above.
After the PC finalizes the proposed Collaboration Plan, it is sent for review and approval by the SCEC Board of Directors
in February. The proposed plan is then sent to the funding agencies for final review. Upon approval by the agencies,
notifications are sent out to the investigators beginning in March. Notifications may be delayed if investigators have
overdue reports on prior SCEC awards.
3.4 Award Procedure
The Southern California Earthquake Center is funded by the National Science Foundation (NSF) and the U.S.
Geological Survey (USGS) through cooperative agreements with the University of Southern California (USC).
Additional funding for the SCEC research program is provided by the Pacific Gas & Electric Company, the Keck
Foundation, geodesy royalty funds, and potentially other sources.
All research awards will be funded as subawards or fixed-priced contracts from the University of Southern California.
Because SCEC5 operates under new cooperative agreements with the NSF and USGS, all 2017 research awards will
be established as new subcontracts between USC and the institution to receive funding. When SCEC award funding
becomes available to investigators will depend on (1) how soon SCEC/USC receives SCEC5 funds from the NSF and
USGS, and (2) how quickly contracts are negotiated between USC and institution to receive funding.
Participant support (workshops, intern project supplement, and travel) award expenditures will be managed through
the master SCEC account at USC.
Funding for Collaborative Proposals may also be delayed or denied if any of the investigators has overdue project
report(s) for prior SCEC award.
3.5 SCEC / USGS-EHP Research Coordination
Earthquake research in Southern California is supported both by SCEC and by the USGS Earthquake Hazards Program
(EHP). EHP's mission is to provide the scientific information and knowledge necessary to reduce deaths, injuries, and
economic losses from earthquakes. Products of this program include timely notifications of earthquake locations, size,
and potential damage, regional and national assessments of earthquakes hazards, and increased understanding of the
cause of earthquakes and their effects. EHP funds research via its External Research Program, as well as work by
USGS staff in its Pasadena (California), Menlo Park (California), Vancouver (Washington), Seattle (Washington), and
Golden (Colorado) offices. The EHP also directly supports SCEC.
SCEC and EHP coordinate research activities through formal means, including USGS membership on the SCEC Board
of Directors, a SCEC/USGS Joint Planning Committee, appointment of a SCEC representative (usually the PC Chair)
to the USGS external research proposal review panel for Southern California, and through a variety of less formal
means. Interested researchers are invited to contact Dr. Rob Graves, EHP coordinator for Southern California, or other
SCEC and EHP staff to discuss opportunities for coordinated research.
The USGS EHP supports a competitive, peer-reviewed, external program of research grants that enlists the talents
and expertise of the academic community, state and local governments, and the private sector. The investigations and
activities supported through the external program are coordinated with and complement the internal USGS program
efforts. This program is divided into six geographical/topical 'regions', including one specifically aimed at Southern
WELCOME
2016 SCEC Annual Meeting page 90
California earthquake research and others aimed at earthquake physics and effects and at probabilistic seismic hazard
assessment (PSHA). The Program invites proposals that assist in achieving EHP goals.
The EHP web page, http://earthquake.usgs.gov/research/external, describes program priorities, projects currently
funded, results from past work, and instructions for submitting proposals. The annual EHP external funding cycle has
different timing than SCEC's, with the USGS RFP due out in February and proposals due in May. Interested PIs are
encouraged to contact the USGS regional or topical coordinators for Southern California, Earthquake Physics and
Effects, and/or National (PSHA) research, as listed under the "Contact Us" tab. The USGS internal earthquake research
program is summarized at http://earthquake.usgs.gov/research/topics.php.
4. Research by Disciplinary Committees
The Center supports disciplinary science through standing committees in Seismology, Tectonic Geodesy, Earthquake
Geology, and Computational Science. They are responsible for disciplinary activities relevant to the SCEC Science
Plan, and they make recommendations to the Science Planning Committee regarding the support of disciplinary
research and infrastructure.
4.1 Seismology
4.1.1 Research Objectives
The Seismology working group gathers data on the range of seismic phenomena observed in southern California,
develops improved techniques for extracting detailed robust information from the data, and integrates the results into
models of velocity structures, source properties and seismic hazard. We solicit proposals that foster innovations in
network deployments and data collection, especially those that fill important observational gaps, real-time research
tools, and data processing. Proposals to develop community products that support one or more SCEC5 goals or include
collaboration with network operators in Southern California are especially encouraged. Proposers should consider the
SCEC resources available including (a) the Southern California Earthquake Data Center (SCEDC) that provides
extensive data on Southern California earthquakes, as well as crustal and fault structure, (b) the network of SCEC
funded borehole instruments that record high quality reference ground motions, and (c) the pool of portable instruments
operated in support of targeted deployments or aftershock response.
4.1.2. Research Strategies
● Enhance and continue operation of the SCEDC and other existing SCEC facilities, particularly the near-real-
time availability of earthquake data and automated access.
● Process network data in real-time to improve estimation of source parameters in relation to faults, short-term
evolution of earthquake sequences, and real-time stress perturbations on major fault segments.
● Enhance and continue to develop earthquake catalogs that include smaller events. Develop improved catalogs
of focal mechanisms and other source properties. Improve locations of important historical earthquakes.
● Enhance or add new capabilities to existing earthquake early warning (EEW) systems or develop new EEW
algorithms. Develop real-time finite source models constrained by seismic and GPS data including robust
uncertainty quantification.
● Advance innovative, practical strategies for densification of seismic instrumentation in Southern California,
including borehole instrumentation; and develop innovative algorithms to utilize data from these networks.
Develop metadata, archival and distribution models for these semi-mobile networks.
● Develop innovative methods to search for unusual signals using combined seismic, GPS, and borehole
strainmeter data; Encourage collaborations with EarthScope or other network operators.
● Investigate near-fault crustal properties, evaluate fault structural complexity, and develop constraints on
crustal structure and state of stress.
● Enhance collaborations, for instance with ANSS, that would augment existing and planned network stations
with downhole and surface instrumentation to assess site response, nonlinear effects, and the ground coupling
of built structures.
● By preliminary design and data collection, seed future passive and active experiments such as dense array
measurements of basin structure, fault zones and large earthquake properties, OBS deployments, and deep
basement borehole studies.
WELCOME
2016 SCEC Annual Meeting page 91
● Investigate whether earthquake properties in southern California have systematic dependencies on properties
of faults, the crust, and anthropogenic activities, which may be used to extract more detailed information from
the available seismic data.
4.1.3 Research Priorities
● Low-cost seismic network data utilization and archiving. Several groups are developing seismic networks
that use low-cost MEMS accelerometers. We seek proposals on innovative algorithms to utilize data from
these networks, develop metadata and archiving models, and make the data and products available to the
user community.
● The shallow crust. Seismic properties in top few kilometers of the crust have strong effects on ground motion,
but are generally not well known. We seek proposals on deriving detailed regional images of seismic velocities
and attenuation coefficients in the shallow crust.
● Tremor and related signals. Tremor has been observed on several faults in California, yet it does not appear
to be ubiquitous. We seek proposals that explore the distribution and source characteristics of tremor in
southern California, and on distinguishing tremor from other sources that may produce similar signals.
● Earthquake directivity. Rupture directivity can have strong influence on ground motion, but it is not clear if
earthquake directivity on given fault sections is systematic or random. We seek proposals on robust
estimations of rupture directivities of a large population of earthquakes in relation to the major faults in southern
California.
● Seismic coupling. The partitioning between seismic and aseismic deformations strongly affects the seismic
potential of faults, but is generally not well known. We seek proposals that develop and implement improved
techniques for estimating the seismic coupling of different fault sections, and for constraining the depth-extent
of seismic faulting in large earthquakes.
● Short-term earthquake predictability. We seek proposals that develop new methods in earthquake statistics
or analyze seismicity catalogs to develop methods for determining short-term (hours to days) earthquake
probability gain.
● Processes and properties in special areas. We seek proposals that use seismic data to improve the
knowledge on structural properties and seismotectonics Southern California, especially those identified as
“Earthquake Gates” (see SAFS for definition).
4.2 Tectonic Geodesy
4.2.1 Research Objectives
The Tectonic Geodesy working group uses mostly geodetic measurements of crustal deformation to understand the
interseismic, coseismic, postseismic, and hydrologic processes associated with the earthquake cycle along the
complex fault network of the Southern San Andreas system. This activity supports several of the SCEC5 science
questions including: How are faults loaded across temporal and spatial scales? What is the role of off-fault inelastic
deformation on strain accumulation, dynamic rupture, and radiated seismic energy? The Tectonic Geodesy group also
plays a role in earthquake monitoring and response, from tracking surface deformation changes that may precede or
accompany induced seismicity, to measuring coseismic displacements and postseismic transients, either in near-real
time as part of earthquake early warning, or as part of a coordinated post-earthquake rapid response.
In addition, the group is tasked with developing a Community Geodetic Model (CGM) for use by the SCEC community
in system-level analyses of earthquake processes over the full range of length and timescales. The CGM is built on the
complementary strengths of temporally dense GPS data and spatially dense InSAR data. Much of the SCEC4 activity
was focused on the assembly and testing of GPS and InSAR data sets for measuring secular motions and comparing
geodetically inferred fault slip rates with geological rates based on paleoseismic studies (e.g. UCERF3). The quality
and quantity of both GPS and InSAR data is rapidly improving to enable a breakthrough in the spatial and temporal
resolution of the CGM. In particular, reprocessing of long GPS time series has provided high accuracy vertical
measurements that reveal a wide range of new hydrologic and tectonic signals. In addition, two new C-band InSAR
satellites (Sentinel-1A and B) are providing highly accurate systematic coverage of the entire SCEC region every 12
days from two look directions. Developing methods to integrate and update these dense spatiotemporal datasets will
be a major task in SCEC5.
WELCOME
2016 SCEC Annual Meeting page 92
4.2.2. Research Strategies
These advances motivate the following strategies:
● Develop vector time series of crustal deformation at ~1 km spatial resolution and better than seasonal temporal
resolution.
● Increase vertical precision of multi-decadal GPS time series.
● Develop methods to robustly estimate spatial and temporal uncertainties in GPS/InSAR/combined time series.
● Provide derived geodetic data products such as strain rate maps, displacement fields, common mode signals,
seasonal signals, and noise assessments.
● Develop methods to constrain the extent of permanent, off-fault deformation, and its contribution to geodetic
fault slip-rate estimates.
● Improve methods for characterizing transient deformation, including episodic aseismic creep, the effects of
surface hydrology and anthropogenic activity, and both short- and long-term postseismic deformation
transients, and for decomposing geodetic time series into secular and transient components.
● Develop and test mechanistic models of the crustal deformation associated with hydrology and/or
anthropogenic activities, with a view to gaining insight into the processes associated with induced seismicity.
● Improve techniques for inferring fault rupture parameters from time-limited geodetic data.
● Work with the initial findings of the Community Rheology Model (CRM) to understand the importance of spatial
variations in rheology on geodetic inversions for fault slip and crustal strain.
● Use these data in combination with other SCEC Community Models, as well as physical models, to address
the major SCEC science questions.
4.2.3 Research Priorities
The major research priorities this year are:
● Community Geodetic Model secular velocities. Produce a consensus secular velocity InSAR product using
the full archive of SAR data (ERS, Envisat, ALOS-1) for the SCEC region. Coordination between participating
groups in this effort is encouraged. Continue to develop methods to improve error characterization and/or
reduction in secular geodetic velocity estimates, e.g. by refined analysis of campaign GPS data in combination
with large-area InSAR analyses, by developing improved noise models, or by improving methods for mitigating
temporally or spatially correlated noise.
● Community Geodetic Model time series. Produce a consensus time series GPS product integrating both
continuous and campaign data. Coordination between participating groups in this effort is encouraged.
Develop methods for processing, updating and integrating InSAR time series in near-real time, using the latest
generation of SAR satellites (e.g. Sentinel-1, ALOS-2). Continue to develop methods for GPS/InSAR time
series integration for the SCEC region, including methods that combine multiple InSAR data sets with different
viewing geometries and temporal sampling rates.
● Research pertaining to other community models. Foster the development of 4D models of the earthquake
cycle that include spatial variations in crustal rheology. Work with other SCEC scientists to develop the
Community Stress Model as well as an improved understanding of how stress varies from the earthquake
cycle timescale to the mountain building timescale (see SDOT).
● GPS data collection and preparation for earthquake rapid response. Collect campaign GPS data in areas
of sparse GPS coverage and poor InSAR correlation. Develop a coordinated rapid response GPS capability,
including updating GPS site coordinates in the vicinity of major faults, in preparation for a major event.
● Community Outreach. Work with the SCEC Communication, Education & Outreach group to highlight the
importance of geodetic measurements for improving earthquake forecasting.
4.3 Earthquake Geology
4.3.1 Research Objectives
The Earthquake Geology working group promotes studies of the geologic record of the Southern California natural
laboratory that advance SCEC science. Its primary focus is on the Late Quaternary record of faulting and ground
motion, including data gathering in response to major earthquakes. Earthquake Geology also fosters research activities
motivated by outstanding seismic hazard issues, understanding of the structural framework and earthquake history of
faults in southern California, and contributes significant information to the statewide Unified Structural Representation
WELCOME
2016 SCEC Annual Meeting page 93
and the Community Rheology Model (CRM). Collaborative proposals that cut across disciplinary boundaries are
encouraged.
4.3.2. Research Strategies
● Paleoseismic documentation of earthquake ages and displacements, with emphasis on long paleoseismic
histories, coordinated slip-per-event and incremental slip-rates, and system-level behavior (seismic
supercycles).
● Gathering well-constrained slip-rates on the southern California fault system, with emphasis on major
structures, along-strike variations, quantification of uncertainties, and comparison with geodetic observations
and fault-loading models.
● Mapping and analysis of fault-zone geometries, mechanical properties, compositions, and fluid histories in
regions where the seismogenic zone, the brittle-ductile transition or the ductile roots of major fault zones have
been exhumed.
● Quantifying variations in fault roughness, complexity, strain localization, and damage in relation to the rupture
propagation processes, including the extent, magnitude, and mechanisms of off-fault deformation.
● Improving the statewide community fault model in areas of inadequate fault representations or where new
data is available, such as using high-resolution topographic data sets to better define fault traces, spatial
uncertainty, and stochastic heterogeneity of fault geometry.
● Validating ground motion prediction through analysis and dating of precariously balanced rocks and other
fragile geomorphic features.
● Developing a geologic framework for the Community Rheology Model that is consistent with existing and
emerging geologic and geophysical observations, and the tectonic history of southern California.
4.3.3 Research Priorities
● Develop a research strategy for the “Earthquake Gates” initiative (see SAFS for definition).
● Foster an interdisciplinary workshop focused on the topic of ‘Beyond Elasticity.’
● Establish a Technical Activity Group (TAG) to develop a geologic framework for the Community Rheology
Model (CRM), as outlined in the 2016 CRM workshop report.
● Document earthquake sources and geomorphic constraints on extreme ground motions in the southern Sierra
Nevada in support of the Central California Seismic Project.
● Revise empirical constraints on fault-rupture hazard (magnitude-displacement scaling) specific to the
California plate boundary fault system.
4.3.4 Geochronology Infrastructure
The shared geochronology infrastructure supports C-14 and cosmogenic dating for SCEC-sponsored research. The
purpose of shared geochronology infrastructure is to allow flexibility in the number and type of dates applied to each
SCEC funded project as investigations proceed. Investigators requesting geochronology support should clearly state
in their proposal an estimate of the number and type of dates required. For C-14, specify if sample preparation will take
place at a location other than the designated laboratory. For cosmogenic dating, investigators are required to arrange
for sample preparation and include these costs in the proposal budget. Investigators are encouraged to contact the
investigators at the collaborating laboratories prior to proposal submission. Currently, SCEC geochronology has
established relationships with the laboratories listed in table below for C-14 and cosmogenic dating.
Investigators may request support for geochronology outside of the infrastructure program for methods not listed here
or if justified on a cost-basis. These requests must be included in the individual proposal budget, and will normally
involve a collaborative proposal with the participating geochronologist. This includes OSL, pIR-IRSL and U-series
dating, each of which are techniques that require that the geochronologist work closely with the project PI. Several
investigators and laboratories, listed below, are available for collaboration on SCEC projects.
WELCOME
2016 SCEC Annual Meeting page 94
Method Laboratory Contact Email
C-14 University of California at Irvine John Southon [email protected]
Lawrence Livermore National Laboratory Tom Guilderson [email protected]
Cosmogenic Lawrence Livermore National Laboratory Susan Zimmerman [email protected]
OSL/pIR-IRSL
University of Cincinnati Lewis Owen [email protected]
Utah State University Tammy Rittenour [email protected]
Desert Research Institute Amanda Keen-Zebert [email protected]
U-Series Berkeley Geochronology Laboratory Warren Sharp [email protected]
Student participation in lab analysis is strongly encouraged by SCEC and by all participating geochronology
laboratories. Please direct questions regarding geochronology infrastructure to the Earthquake Geology group leader,
Mike Oskin ([email protected]).
Data Reporting Requirements. Funded investigators are required to provide full reporting of their geochronology
samples, including raw data, interpreted age, and geographic/stratigraphic/geomorphic context (what was dated?)
before samples are submitted for analysis.
4.4 Computational Science (CS)
4.4.1 Research Objectives
The Computational Science working group promotes the use of advanced numerical modeling techniques, data
intensive (big-data) computing, and high performance computing (HPC) to address the emerging needs of SCEC users
and the application community on a variety of computer systems, including HPC platforms. The group works with SCEC
scientists across a wide range of topics to take advantage of rapidly changing computer architectures and algorithms.
We engage and coordinate with national HPC labs, centers, and service providers in crosscutting efforts to enable
large-scale and data-intensive computing milestones. The group encourages research using national supercomputing
resources, and supports students from both geoscience and computer science backgrounds to develop their skills in
the area. Projects listing Computational Science as their primary area should involve significant software-based
processing, big-data synthesis, or HPC in some way. Proposed research with the potential to be transferred to data
intensive and HPC applications should make this explicit and include Computational Science as a secondary focus
area. Research utilizing standard desktop computing should list the most relevant non-Computational Science working
group as the primary area.
Computational Requirements. If your proposed research requires substantial SCEC computing resources or
allocations, your proposal to SCEC should detail the computational requirements and include the following information:
1. The scientific goal of your computational research.
2. The scientific software you plan to use or develop.
3. A list of computations you plan to run.
4. The estimated computing time that will be required.
5. The computer resources you plan to use to perform your simulations.
Investigators on proposals that anticipate use of SCEC computational resources and/or help from SCEC software
developers should consult with SCEC IT leadership for budget time support estimates and coordination planning.
Note that researchers may request startup allocations from NSF’s Extreme Science and Engineering Discovery
Environment (XSEDE) at https://www.xsede.org/allocations.
4.4.2. Research Strategies
● Reengineer and optimize HPC codes, required to reach SCEC research goals, for CPU-based and/or GPU-
based supercomputers, with emphasis on issues such as performance, portability, interoperability, power
efficiency, and reliability.
● Develop novel algorithms and implementation of more realistic models for earthquake simulation, particularly
those that improve efficiency and accuracy or expand the class of problems that can be solved (see SCEC5
Thematic Areas).
● Optimize earthquake cycle simulators that can resolve fault processes across the range of scales required to
investigate stress-mediated fault interaction, including those caused by dynamic wave propagation, or that
WELCOME
2016 SCEC Annual Meeting page 95
combine coseismic dynamic rupture and multi-cycle simulators; generate synthetic seismicity catalogs; and
assess the viability of earthquake rupture forecasts.
● Develop tools and algorithms for uncertainty quantification in large-scale inversion and forward-modeling
studies, for managing I/O, data repositories, workflow, advanced seismic data format, visualization, and end-
to-end approaches.
● Develop data-intensive computing tools, including but not limited to, InSAR and geodesy, lidar and structure-
from-motion, 3D tomography, cross-correlation algorithms used in ambient noise seismology, and other signal
processing techniques used, for example, to search for tectonic tremor and repeating events.
● Develop data-structure libraries and algorithms that contribute to integration of SCEC community models
(CXM) with HPC applications for modeling and simulation; with emphasis in accelerating model creation
(meshing and gridding), visualization of models, flexible integration of new datasets; and community software
for CXM development.
4.4.3 Research Priorities
● Seismic wave propagation
○ Develop tools and implement procedures to validate SCEC community fault and seismic velocity
models as applied in inverse and forward problems.
○ Develop and improve existing software tools and algorithms with application to HPC that accelerates
and advances high-frequency simulation methods, while contributing to solve standing research
problems as stated in the Ground Motions working group priorities.
○ Develop software tools and algorithms to incorporate more realistic constitutive material models and
material model representation with application in wave propagation HPC codes. Examples include,
but are not limited to, developing efficient tools and algorithms for modeling inelastic deformation,
scattering by small-scale heterogeneities, and topography.
● Tomography
○ Develop new or adapt existing forward modeling software to solve full 3D tomography (F3DT)
problems using HPC resources. Computational Science research in this area should help diversify
current dependency on existing software and provide alternatives for SGT and F3DT verification, as
done in dynamic rupture and forward ground motion simulation.
○ Develop efficient and sustainable computational procedures to facilitate the assimilation of regional
waveform data in the SCEC community velocity models; and integrate F3DT and inversion results
into existing community (velocity) models.
● Rupture dynamics
○ Develop computational codes that incorporate more sophisticated and realistic descriptions of fault
weakening processes, complex fault geometry, and material heterogeneity and inelastic response in
large-scale earthquake simulations, ideally with minimal impact on computational efficiency. This
might require novel numerical approaches like adaptive mesh refinement.
○ Develop computational codes that merge single event rupture models with earthquake cycle models.
● Scenario earthquake modeling
○ Develop software processing tools such as workflows that can facilitate modeling a suite of scenario
ruptures, incorporating material properties and fault geometries from SCEC community models, and
investigate the sensitivity of simulations to models and modeling parameters.
○ Isolate causes of amplified ground motion using adjoint-based sensitivity methods.
● Data-intensive computing
○ Develop computational tools for advanced signal processing algorithms, such as those used in
ambient noise seismology and tomography as well as InSAR and other forms of geodesy.
○ Integrate Big Data analytics techniques involving software stacks such as Hadoop, fault recovery,
data format, generation, partitioning, abstraction, and mining.
● Engineering applications
○ Develop computational tools to investigate the implications of ground motion simulation results by
integrating observed and simulated ground motions with engineering-based building response
models; and to validate the results by comparison to observed building responses.
○ Develop and advance existing computational platforms that will facilitate end-to-end “ruptures-to-
rafters” modeling capabilities that can help transform earthquake risk management into a
cyberinfrastructure science and engineering discipline.
WELCOME
2016 SCEC Annual Meeting page 96
5. Research by Interdisciplinary Working Groups
SCEC coordinates earthquake system science through interdisciplinary working groups including: Fault and Rupture
Mechanics (FARM), Stress and Deformation Over Time (SDOT), Earthquake Forecasting and Predictability (EFP), and
Ground Motions. The Southern San Andreas Fault Evaluation (SoSAFE) group has evolved into the San Andreas Fault
System (SAFS) working group for SCEC5. This group coordinates research on the San Andreas and the San Jacinto
master faults. Also new in SCEC5 is the SCEC Community Models (CXM) working group, focused on developing,
refining and integrating community models that describe a wide range of features of the southern California lithosphere
and asthenosphere. Seismic hazard and risk analysis research continues to be coordinated by the Earthquake
Engineering Implementation Interface working group. Their activities include educational as well as research
partnerships with practicing engineers, geotechnical consultants, building officials, emergency managers, financial
institutions, and insurers.
5.1 Fault and Rupture Mechanics (FARM)
5.1.1 Research Objectives
The Fault and Rupture Mechanics (FARM) working group develops (1) constraints on the properties, conditions and
physical processes that control faulting in the lithosphere over the entire range of pre-, co-, and post-seismic periods
using field, laboratory, and theoretical studies; and (2) physics-based fault models applicable to various spatial and
temporal scales, such as nucleation, propagation and arrest of dynamic rupture or long-term earthquake sequence
simulations. This fundamental research aims to develop physics-based understanding of earthquakes in the Southern
California fault system and contribute to SCEC hazards estimates such as the Uniform California Earthquake Rupture
Forecast (UCERF) and physics-based ground motion predictions.
5.1.2. Research Strategies
Observational constraints on earthquake physics come from a range of sources, including seismology, geodesy, field
geology, borehole geophysics, heat flow, hydrology, and gravity. Insights into earthquake physics are obtained by
targeted laboratory fault and rock mechanics experiments and theoretical studies. Numerical modeling is used for
understanding, analyzing, and relating theories, laboratory findings, and observations as well as exploring the
implications of the findings. FARM supports fundamental research using these and related approaches. FARM research
strategies include:
● Field, laboratory, and theoretical studies to determine spatial (including depth-dependent) and temporal
variations in evolving fault strength and the effect of various relevant factors such as temperature, composition,
pore pressure, degree of shear localization, and damage, including the variations and heterogeneity of
relevant fault and rock properties.
● Theoretical and laboratory studies of nucleation, propagation, and arrest of seismic and aseismic fault slip.
● Seismological, geodetic, geophysical, laboratory and theoretical determinations of earthquake triggering,
including triggering by static and dynamic stressing, fluid injection (induced earthquakes), and aseismic
deformation.
● Geologic descriptions of fault complexity and shear zone structure, their relation to fault-scale and system-
scale structural complexity, and representations suitable for inclusion in physics-based models.
● Characterization of fault damage zones and their evolution over both seismic and interseismic periods using
seismological, geodetic, geologic, laboratory, and numerical experiments.
● Inferences and measurements of fault zone pore pressure, fluid flow, and their temporal and spatial variation.
● Development of improved numerical approaches to interrogate various temporal and spatial scales of faulting,
including long-terms simulations of fault slip that incorporate inertial effects during earthquakes and
geologic/fault system scale earthquake simulations (simulators) for the next generation of seismic hazard
estimates ('beyond PSHA').
5.1.3 Research Priorities
● Constrain how absolute stress, fault strength and rheology vary with depth on faults.
● Determine the mechanisms dominant in co-seismic (dynamic) fault resistance, including the relative
importance of various potential dynamic weakening mechanisms and off-fault processes, and their
compatibility with observational constraints such as stress drops and temperature measurements.
WELCOME
2016 SCEC Annual Meeting page 97
● Investigate the effect of fine-scale processes on the nucleation and dynamic rupture of large earthquakes and
the resulting ground shaking, including whether the fine-scale processes can be suitably coarse-grained at
the system scale.
● Quantify stress heterogeneity on faults at different spatial scales and determine whether it correlates with
asperities and other geometric complexities.
● Investigate the relation between material, geometrical, and dynamic (deformation-induced) on- and off-fault
heterogeneity, its effect on rupture initiation, propagation, and arrest, and implications for radiated energy, slip
and rupture speed distributions, their scaling, and ground motion.
● Investigate how inelastic strain associated with fault roughness and discontinuities influences rupture
propagation, seismic radiation, and scaling of earthquake source parameters.
● Determine how damage zones, crack healing and cementation, fault zone mineralogy, and off-fault plasticity
govern strain localization, slip stability, interseismic strength recovery, and rupture propagation.
● Determine how much off-fault inelasticity contributes to strain accumulation and what fraction of strain is
relaxed by aseismic processes.
● Describe how fault complexity and inelastic deformation interact to determine the probability of rupture
propagation through structural complexities.
● Assess how shear resistance and energy dissipation depend on the maturity of the fault system, and how
these are expressed geologically.
● Constrain the active geometry and rheology of the vicinity of brittle-ductile transition - including frictional sliding
stability transition zone - and ductile roots of fault zones, and determine their implications for depth limits of
large earthquakes, overlap of seismic and aseismic slip, geodetic estimates of fault locking, and transition
from frictional sliding to visco-plastic flow.
● Study the mechanical and chemical effects of fluid flow, both natural and anthropogenic, on faulting and
earthquake occurrence, and how they vary throughout the earthquake cycle.
● Determine how seismic and aseismic deformation processes interact, and how that interaction affects long-
term fault behavior, by developing numerical simulations that incorporate both seismic and interseismic
periods and a combination of relevant physical factors, e.g., establishing the long-term effect of off-fault
damage during earthquakes and off-fault healing during the interseismic periods; exploring how slow slip and
microseismicity redistributes stress for the following large events, and determining how large events can
rupture into interseismically stable fault regions due to co-seismic weakening.
● Study the implications of earthquake physics findings on earthquake hazard by developing physics-based
long-term simulations of earthquake sequences on fault systems (earthquake simulators).
● Exploit anthropogenic (induced) seismicity as experiments in earthquake physics and earthquake
predictability.
5.2 Stress and Deformation Over Time (SDOT)
5.2.1 Research Objectives
The focus of Stress and Deformation Over Time (SDOT) working group is to improve our understanding of how faults
in the crust are loaded in the context of the wider lithospheric system. SDOT studies lithospheric processes on
timescales from tens of millions of years to tens of years using the structure, geological history, and physical state of
the southern California lithosphere as a natural laboratory. One objective is to characterize the present-day state of
stress and deformation on crustal-scale faults and the lithosphere as a whole, and to tie this stress state to the long-
term evolution of the lithospheric architecture through geodynamic modeling. Another central goal is to contribute to
the development of a physics-based, probabilistic seismic hazard analysis for southern California by developing and
applying system-wide deformation models.
5.2.2. Research Strategies
Addressing the SDOT research objectives requires a better understanding of fundamental quantities including
lithospheric driving forces, the relevant rock rheology for processes acting over a wide range of time scales, fault
constitutive laws, and the spatial distribution of absolute deviatoric stress. Tied in with this is a quest for better structural
constraints, such as on density, rock type, Moho depths, thickness of the seismogenic layer, the geometry of
lithosphere-asthenosphere boundary, as well as basin depths, rock type, temperature, water content, and seismic
WELCOME
2016 SCEC Annual Meeting page 98
velocity and anisotropy. These quantities are constrained by a wide range of observables from disciplines including
geodesy, geology, and geophysics.
Specific SDOT strategies include:
● Seismological imaging of the crust, lithosphere and upper mantle using interface and transmission methods.
● Examinations of geologic inheritance and evolution, on faults and off, and its relation to the three-dimensional
(3D) structure and physical properties of the present-day crust and lithosphere.
● Development of models of interseismic, earthquake cycle and long-term deformation.
● Development of models using approaches that may include analytical or semi-analytical methods, spectral
approaches, boundary, finite, or distinct element methods.
5.2.3 Research Priorities
● Characterize the 3D distribution of isotropic and anisotropic wave speed variations. Assemble 3D lithological
models of the crust, lithosphere, and mantle based on active- and passive-source seismic data, potential field
data, and surface geology.
● Contribute to the development of the geologic framework for the Community Rheology Model (CRM).
● Advance research into averaging, simplification, and coarse-graining approaches in numerical models across
spatio-temporal scales, addressing questions such as the appropriate scale for capturing fault interactions,
the adequate representation of frictional behavior and dynamic processes in long-term interaction models,
fault roughness, structure, complexity and uncertainty.
● Develop models to estimate slip rates on southern California faults, fault geometries at depth, and spatial
distribution of slip or moment deficits on faults. Incorporate rheological and geometric complexities in such
models and exploration of mechanical averaging properties. Assess potential discrepancies of models based
on geodetic, geologic, and seismic data.
● Develop deformation models (fault slip rates and locking depths, off-fault deformation rates) in support of
earthquake rupture forecasting.
● Determine the spatial scales at which tectonic block models (compared to continuum models) provide
descriptions of fault system deformation that are useful for earthquake forecasting.
● Develop interseismic and long-term deformation models that incorporate inelastic deformation. Develop that
can predict the relative proportions of elastic, recoverable strain associated with the earthquake cycle and
permanent, distributed strain in the crust.
● Apply stress and deformation measurements at various time scales for hypothesis testing of issues pertaining
to postseismic deformation, fault friction, rheology of the lithosphere, seismic efficiency, the heat flow paradox,
stress and strain transients, and fault system evolution. Improved constraints on the active geometry and
rheology of ductile roots of fault zones.
● Contribute to the Community Stress Model (CSM). Development of spatio-temporal (4-D) representations of
the stress tensor in the southern California lithosphere using diverse stress constraints (e.g. from borehole or
anisotropy measurements) and geodynamic models of stress.
● Contribute to the Community Thermal Model (CTM) and Community Rheology Model (CRM). Test provisional
rheology models with numerical models and geodetic, seismic, and geologic data.
● Develop vertical deformation models (interseismic and multi-earthquake cycle) that incorporate improved
vertical constraints from the Community Geodetic Model (CGM). Improve analyses of model sensitivity to non-
tectonic vertical motion signals.
5.3 Earthquake Forecasting and Predictability (EFP)
5.3.1 Research Objectives
The Earthquake Forecasting and Predictability (EFP) working group coordinates five broad types of research projects:
(1) the development of earthquake forecast methods, (2) the development of testing methodologies for evaluating the
performance of earthquake forecasts, (3) expanding fundamental physical or statistical knowledge of earthquake
behavior that may be relevant for forecasting earthquakes, (4) the development and use of earthquake simulators to
understand predictability in complex fault networks, and (5) fundamental understanding of the limits of earthquake
predictability.
WELCOME
2016 SCEC Annual Meeting page 99
5.3.2. Research Strategies
We seek proposals that will increase our understanding of how earthquakes might be forecasted, to what extent and
precision earthquakes are predictable, and what is the physical basis for earthquake predictability. Proposals of any
type that can assist in this goal will be considered. In order to increase the amount of analyzed data, so as to decrease
the time required to learn about predictability, proposals are welcome that deal with global data sets and/or include
international collaborations.
For research strategies that plan to utilize the Collaboratory for the Study of Earthquake Predictability (CSEP), see
Section 6.2 to learn of its capabilities. Successful investigators proposing to utilize CSEP would be funded via core
SCEC funds to adapt their prediction methodologies to the CSEP framework, to transfer codes to the externally
accessible CSEP computers, and to be sure they function there as intended. Subsequently, the codes would be moved
to the identical externally inaccessible CSEP computers by SCEC IT staff who will conduct tests against a variety of
data as outlined in the CSEP description.
5.3.3 Research Priorities
Determine the spatial scales at which tectonic block models (compared to continuum models) provide descriptions of
fault-system deformation that are useful for earthquake forecasting.
Describe how fault complexity and inelastic deformation interact to determine the probability of rupture propagation
through structural complexities, and determine how model-based hypotheses about these interactions can be tested
by the observations of accumulated slip and paleoseismic chronologies.
Constrain the extent of permanent, off-fault deformation, and its contribution to geologic and geodetic fault slip-rate
estimates.
Study the mechanical and chemical effects of fluid flows, both natural and anthropogenic, on faulting and earthquake
occurrence, and how they vary throughout the earthquake cycle.
Develop earthquake simulators that encode the current understanding of earthquake predictability.
Place useful geologic bounds on the character and frequency of multi-segment and multi-fault ruptures of extreme
magnitude.
Assess the limitations of long-term earthquake rupture forecasts by combining patterns of earthquake occurrence and
strain accumulation with neotectonic and paleoseismic observations of the last millennium.
Test the hypothesis that “seismic supercycles’’ seen in earthquake simulators actually exist in nature and explore the
implications for earthquake predictability.
Exploit anthropogenic (induced) seismicity as experiments in earthquake predictability.
● Support the development of statistical or physics-based real-time earthquake forecasts.
● Utilize and/or evaluate the significance of earthquake-cycle simulator results. See WGCEP and CSEP
sections for more details.
● Study how to properly characterize and estimate various earthquake-related statistical relationships (including
the magnitude distribution, Omori law, aftershock productivity, etc.).
● Focus on understanding patterns of seismicity in time and space, as long as they are aimed toward
understanding the physical basis of earthquake predictability.
● Develop useful measurement/testing methodology that could be incorporated in the CSEP evaluations,
including those that address how to deal with observational errors and other uncertainties in data sets.
● Develop approaches to test the validity of the characteristic earthquake vs. Gutenberg-Richter earthquake
models as they are used in seismic hazard analysis.
5.4 Ground Motions (GM)
5.4.1 Research Objectives
The primary goal of the Ground Motions (GM) interdisciplinary working group is to study the characteristics of ground
motion data, understand the complex mechanisms that give rise to these characteristics, use this understanding to
develop and implement physics-based simulation methodologies that can predict strong-motion broadband (up to 10
WELCOME
2016 SCEC Annual Meeting page 100
Hz) waveforms and their effects on the natural geologic environment, and validate the numerical models using ground
motion data and their statistics.
An important focus of SCEC5 is to investigate the role of nonlinear material deformation and material failure on
simulated ground motions. In the context of this working group, we seek to understand and simulate the conditions
under which nonlinear effects reduce the ground surface shaking but cause permanent deformations that can be
damaging to distributed infrastructure systems; the conditions which lead to site amplification due to impedance
contrasts between stiff and soft crustal layers accentuated by material yielding; and the more dramatic effects of ground
failure such as liquefaction and landslides.
Another important focus for SCEC5 is development of methodologies for validation of ground-motion simulations
against observed data, and implementation and testing of those methods. The ground-motion simulations should be
valid across a range of magnitudes, distances, and frequency bands, but especially for large magnitudes at close
distances.
Both source and path characterization play a vital role in ground-motion prediction and are important areas of research
for the group; by path we refer both to the deeper layers of the crust (rock) and to the near-surface soft sedimentary
deposits (soil) of the numerous basins in Southern California. Although there exist numerous sophisticated constitutive
soil and rock models that have been developed using results of extensive site-specific field and laboratory testing, they
cannot be used in regional-scale physics-based simulations because measurements of their input parameters are
scarcely available at the simulation target resolution - if at all. Therefore, to incorporate nonlinear effects in physics-
based scenario simulations, the GM working group seeks the development and implementation of simplified constitutive
models based on very few physical parameters; tailored to the target simulated phenomena (e.g. amplification vs
liquefaction); and accompanied by empirical correlations that translate available material properties or proxies into input
parameters for these constitutive models.
Note that GM working group also shares interests with the Earthquake Engineering Implementation Interface (EEII),
Computational Science (CS), and Special Projects and consult those sections for additional GM-related research
priorities.
5.4.2. Research Strategies
● Analyze ground motion data to understand the physical mechanisms that control the observed characteristics.
● Develop and implement ground-motion simulations and simplified constitutive models to represent a variety
of magnitude and distance ranges, as well as a multitude of secondary effects.
● Develop high-resolution, 3D rheology models (velocity, anelastic attenuation, nonlinear properties) of the near-
surface crustal layers for integration in physics-based nonlinear wave propagation simulations.
● Validate ground motion simulations through development and testing of algorithms that trace the predictability
and uncertainties of the simulated ground motions across frequency ranges).
● Continue development of the SCEC Broadband Platform.
● Coordinate efforts across the working groups, possibly via the formation of Technical Activity Groups (TAGs).
Example of a TAG that would benefit the GM group would focus on the modeling, integration and simulation
of nonlinear phenomena in the near-surface soft sediments (effort cuts across CS, EEII, CXM and GM).
5.4.3 Research Priorities
● Gather and develop novel data sets to be used as ground motion data for application to any of the research
priorities. This may include, but certainly not limited to, using small magnitude earthquake data, ambient noise
data or strain meter data.
● Use observed ground-motion data to infer physical understanding of the controls on ground motion, to aid in
the development of ground-motion simulations.
● Develop ground motion simulation methods to capture observed ground-motion phenomena, especially to
account for the paucity of recordings in the near-field, where the motions are strong and inelastic effects may
be large. These models should be tailored to the target simulated phenomena (e.g. site amplification, basin
effects, topography effects, liquefaction), and accompanied by empirical correlations that translate available
material properties or proxies into input parameters for these constitutive models.
● Determine the relative roles of fault geometry, heterogeneous frictional resistance, wavefield scattering,
intrinsic attenuation, and near-surface nonlinearities in controlling ground motions.
WELCOME
2016 SCEC Annual Meeting page 101
● Develop constitutive models to capture nonlinear phenomena such as permanent ground deformation and
earthquake triggered ground failure phenomena in physics-based simulations. The models should be
formulated on the basis of a few input parameters feasible to obtain (or correlate with available measurements)
on a regional scale. Quantify the epistemic and parametric uncertainties of these models; and develop
protocols on how to map material proxies (such as Vs30, or velocity profiles where available) and empirical
estimates of the nonlinear soil and rock parameters (e.g., friction angle as a function of confining pressure) on
the input parameters. This effort should be coordinated with the research priorities and activities of the SCEC
Community Models (CXM) especially the velocity and rheology models; the Fault and Rupture Mechanics
(FARM) group with respect to the development and implementation of consistent material models for fault-
zone and off-fault plasticity of rocks; and the Computational Science (CS) group for the computationally
efficient integration of these models in large-scale simulations.
● Develop high-resolution, 3D rheology models (velocity, anelastic attenuation, nonlinear properties) of the near-
surface soft sedimentary layers of the crust for integration in physics-based nonlinear wave propagation
simulations of broadband waveforms (<10Hz). Use available borehole measurements, near-surface material
stiffness proxies (e.g., Vs30, topography) and/or empirical correlations to estimate input parameters necessary
for the physics-based nonlinear ground motion simulations. This effort should be coordinated with the research
priorities and activities of the SCEC Community Models (CXM), to ensure that the sedimentary models are
properly constrained by the rheological properties of the deeper Community Velocity Model structure
(including the 3D geometry of sedimentary basins), and avoid spurious reflections from fictitious impedance
contrasts that could emerge from improper integration of the sedimentary layers with the underlying rock
basement.
● Develop statistical models of the sedimentary layers for stochastic media realizations, intended to capture
scattering attenuation for the SCEC Broadband Platform (BBP) and for 3D physics-based simulations.
Consider the depositional environment and the differences between lateral and vertical heterogeneities.
Compare simulated ground motions from rough faults propagating through realistic heterogeneous media to
simulated ground motions representing scattering via anelastic frequency dependent attenuation and to
historical data.
● Develop, in collaboration with the Computational Science (CS) group, computationally efficient methodologies
of integrating topographic amplification in 3D simulations, and study the effects of scattering and focusing by
the geometric irregularities, as well as the coupling between wave amplification effects in the near-surface
damaged rock layers and topography effects.
● Explore approaches to represent the effects of non-linearity through computationally efficient methodologies
based on elasticity, especially for SCEC products that entail very large numbers of simulations such as
Cybershake. Develop approaches appropriate for ground-motion simulations for anticipated large events that
are suitable for probabilistic seismic hazard and risk analysis, and potentially distinguish between near-field
and far-field nonlinear effects.
● Conduct systematic verification (comparison against theoretical predictions) of the simulation methodologies
with the objective to develop robust and transparent simulation capabilities that incorporate consistent and
accurate representations of the earthquake source and three-dimensional velocity structure.
● Conduct systematic validation of ground-motion simulation methodologies to appropriate observed historical
ground-motion data, with the objective to develop robust and transparent simulation capabilities that
incorporate consistent and accurate representations of the earthquake source and three-dimensional velocity
structure. Trace the modeling and parametric uncertainties of the simulated ground motions. Validate
individual components of ground-motion predictions, such as the source, site or path components. Conduct
validation across frequency ranges (each representative of the scale of the model input parameters and source
models). Compare synthetic ground motions from deterministic and stochastic approaches to data for
overlapping bandwidths. Validate specific elements of the complex physics-based simulation models to
identify topics where improvement is necessary.
● Validate ground-motion simulations against GMPEs (or statistical trends in GMPEs) and engineering metrics.
● Incorporate new effects and features in the SCEC Broadband Platform, such as multi-segment rupture;
nonlinear amplification factors; spatial variability of crustal properties; 3D basins and their on ensemble
averaged long-period ground motions (e.g., by comparing ensemble averages of long-period (<~1Hz) ground
motions computed in 1D and 3D crustal models for events included in the GMSV). Develop and implement
WELCOME
2016 SCEC Annual Meeting page 102
methods for computing and storing 3D Green's functions (GFs) for use in the SCEC BBP. Proposals for both
source- and site-based GFs are solicited (see CXM section for related efforts).
● Develop frequency-dependent amplification factors for the SCEC BBP (preferably based on Fourier-spectral
ratios, that capture both amplitude and phase), and parameterized in terms of simple site parameters such as
Vs30, basin depth, and/or the near surface soil-velocity gradient. The proposed complex amplification factors
should be developed on the basis of an exhaustive set of analyses with 1D site response models that capture
soil behavior in the linear elastic, nonlinear elastic and and plastic strain regimes.
5.5 San Andreas Fault System (SAFS)
5.5.1 Research Objectives
The San Andreas Fault System (SAFS) group will develop projects within SCEC that are focused on the occurrence of
large earthquakes on the San Andreas Fault system, seeking (1) projects that continue data collection on the timing
and size of earthquakes on the SAF and (2) projects that investigate the features of the fault system that may halt or
permit continued rupture. This second class of projects falls under a new SCEC5 initiative called “Earthquake Gates.”
Like the special fault study areas of SCEC4, Earthquake Gates projects will bring together groups of collaborators
across multiple disciplines but the focus of the projects will be to investigate the features that promote or arrest large
earthquakes on the San Andreas Fault System. For example, studying features that may conditionally arrest ruptures
so that at some times the gate is ‘closed’ while at other times the gate is ‘open’. Understanding the character of such
features will help us better forecast earthquake behavior.
5.5.2. Research Strategies
Research foci in the SAFS group will include basic studies on the occurrence of large earthquakes on the San Andreas
Fault System and collaborative proposals under the Earthquake Gates initiative. Earthquake Gates (EG) projects are
expected to be organized around multi-year and multi-investigator, collaborative research plans. To establish
Earthquake Gates projects, a five page, “Earthquake Gates Proposal” must be submitted for review by SCEC Board
and PC. The general EG proposal will establish the research questions to be addressed, methodologies and goals,
and a timeline for the project. The scope should be achievable within the five year period of SCEC5.
To develop the Earthquake Gate proposal for 2017, investigators are encouraged to either:
(1) Propose to hold a workshop in which individual investigators will meet and develop an Earthquake Gates proposal.
The workshop proposal should (a) follow standard workshop proposal guidelines (see Proposal Categories) and (b)
include a mechanism for notifying the community of the opportunity. Field trips may be included with this proposal.
OR
(2) Develop an Earthquake Gates proposal in advance or without a workshop. In this case, leaders of the Earthquake
Gates proposal should reach out to the community to develop the collaborative research plan. Individual PIs will submit
proposals—identified as part of an Earthquake Gates project—that details the methods and scope of their part of the
project and how it fits into the overall EG proposal.
5.5.3 Research Priorities
● Basic research on the occurrence of large earthquakes on the San Andreas Fault System.
● Earthquake Gates projects should focus on either
○ A subregion within Southern California in which subsurface data such as paleoseismic or geophysical
data can be combined with modeling approaches to study the hazards and earthquake behavior in
that region, or
○ An idealized type of fault system or geometry that has examples within Southern California.
The goal is to understand in greater detail the attributes such as fault geometry, rupture direction, prior earthquake
history, as well as off-fault rheology to study if earthquakes halt or continue to grow.
5.6 SCEC Community Models (CXM)
5.6.1 Research Objectives
The SCEC CXM working group develops, refines and integrates community models describing a wide range of features
of the southern California lithosphere and asthenosphere. These features include: elastic and attenuation properties
WELCOME
2016 SCEC Annual Meeting page 103
(Community Velocity Model, CVM), temperature (Community Thermal Model, CTM), rheology (Community Rheology
Model, CRM), stress and stressing rate (Community Stress Model, CSM), deformation rate (Community Geodetic
Model, CGM), and fault geometry (Community Fault Model, CFM). Our goal is to provide an internally consistent suite
of models that can be used to simulate the seismic behavior in southern California.
SCEC5 research goals involve continued refinement of existing community models (CFM, CVM, CSM, CGM),
development of new community models (CTM and CRM), and integration of the models into a self-consistent suite.
Objectives also include quantification of uncertainties and development of techniques for propagating uncertainties
from observations through community model development to simulation predictions.
5.6.2. Research Strategies
● Develop and apply inversion techniques to populate and refine the community models.
● Collect additional observations to improve resolution of a community model and/or resolve discrepancies
among competing models.
● Develop viable alternative community models that facilitate representation of the epistemic uncertainty.
● Develop methods to characterize uncertainty in each of the community models.
● Validate individual community models and verify consistency across multiple community models (e.g.,
consistency of stress predictions from the CTM and CRM with the CSM).
● Use community models in simulations to forecast behavior, including estimates of the uncertainties in
predicted values.
● Expand community participation in model development, validation, and application through workshops,
tutorials, and participation in and/or collaboration with related efforts (e.g., EarthCube).
5.6.3 Research Priorities
● Science integration. Convene a workshop focused on guiding community model development towards self-
consistent and well-integrated community models.
● Information technology integration. In collaboration with the Community Modeling Environment (CME) and
the CXM developer- and user-communities, develop detailed documentation defining standard user-
interface(s) and storage formats for the community models. Design goals include: ease of use, computational
efficiency, and common interfaces and storage for each class of model (e.g., grid-based or object-based). The
documentation should include complete specification of user-interfaces as well as appropriate use cases for
each community model. Emphasis is on defining long-lasting, effective standards over an initial
implementation of the design.
● Training. Provide virtual or in-person hands-on training in how to use the existing community models in
research applications. Training should be archived for on-demand access.
● Small scale features. Develop methods to represent smaller scale features in the CXMs, such as stochastic
variations in elastic properties, attenuation, stress, temperature, rheology, and fault strike and dip.
● Validation. Develop and apply procedures (i.e., goodness-of-fit measures) for evaluating updated CXMs
against observations (e.g., seismic waveforms, gravity, etc) to discriminate among alternatives and quantify
model uncertainties.
● Community Fault Model (CFM)
○ Improve and evaluate the CFM, placing emphasis on defining the geometry of major faults that are
incompletely, or inaccurately, represented in the current model, and on faults of particular concern,
such as those that are located close to critical facilities.
○ Refine representations of the linkages among major fault systems.
○ Generate maps of geologic surfaces compatible with the CFM that may serve as strain markers in
crustal deformation modeling and/or property boundaries in future iterations.
○ Collaborate with developers of related data products to improve consistency in the organization and
naming scheme (e.g., USGS Fault and Fold database).
● Community Velocity Model (CVM)
○ Integrate new data (especially the Salton Sea Imaging Project) into the existing CVMs with validation
of improvements in the CVMs for ground-motion prediction.
○ Quantify uncertainty in the CVM structure and its impact on simulated ground motions.
WELCOME
2016 SCEC Annual Meeting page 104
● Community Stress Model (CSM)
○ Populate the CSM and assess sensitivity of stress and deformation patterns to parameter variations
(what level of detail is needed in the CRM/CTM?).
○ Compile diverse constraints on stress (e.g. from borehole or anisotropy measurements) and evaluate
the accuracy of the CSM.
● Community Geodetic Model (CGM)
○ Extend the existing CGM to include time dependence and vertical motion.
○ Create realistic measures of uncertainty for the horizontal and vertical motions.
○ Develop products such as strain rate maps, displacement fields, common mode and seasonal
signals, and noise assessments. Implement standardized procedures for generating these products.
● Community Thermal Model (CTM)
○ Deliver a preliminary CTM that provides temperatures throughout the southern California lithosphere
and asthenosphere consistent with radiogenic heat production.
○ Improve/interpolate surface heat flow maps. Search for additional heat flow and thermal property
data in areas with poor coverage.
● Community Rheology Model (CRM)
○ Develop a geologic framework for the Community Rheology Model, as outlined in the 2016 CRM
workshop report.
○ Construct a provisional rheology model based upon a simplified geologic framework. The model
should provide constitutive relations for ductile, plastic and brittle material in rock and in shear zones.
○ Develop mixing laws for polymineralic rocks of the CRM.
5.7 Earthquake Engineering Implementation Interface (EEII)
5.7.1 Research Objectives
The purpose of the Earthquake Engineering Implementation Interface (EEII) is to create and maintain collaborations
with research and practicing engineers. These activities may include ground motion simulation validation, as well as
the end-to-end analysis of structures and infrastructure systems. Our goal of impacting engineering practice and large-
scale risk assessments requires partnerships with the engineering and risk-modeling communities, which motivates the
following activities.
5.7.2. Research Strategies
Example strategies to achieve these objectives include:
● Perform engineering and risk analysis using SCEC research products related to hazards and ground motions,
in order to determine the impact of research insights on engineering decisions, and sensitivity of engineering-
related result to parameters in the science models.
● Develop tools and approaches that facilitate the transfer or SCEC science products to the research
community.
● Form groups to reach consensus on methods to validate and utilize simulated ground motions, simulation-
based hazard maps, and other SCEC science products of relevance to engineers and risk analysts.
5.7.3 Research Priorities
Ground motion simulation validation and utilization
● Develop, coordinate and vet methods for validating simulations of ground motions for engineering use.
● Demonstrate ground motion simulation validation methodologies with existing simulated ground motions.
● Develop methodologies to validate and use SCEC CyberShake ground motion simulations for developing
probabilistic and deterministic hazard maps for building codes and other engineering applications seismic
hazard applications. Investigations of observed versus simulated region-specific path effects for small-
magnitude earthquakes in Southern California are encouraged.
● Develop data, products or tools that enable physics-based hazard calculations or ground motions to be utilized
by engineers and risk modelers.
● Identify or demonstrate links between ground motion metrics or structural response parameters and ground
motion simulation features, such that simulation algorithms might be improved to better represent ground
motion features of interest.
WELCOME
2016 SCEC Annual Meeting page 105
● Quantify and evaluate spatial variation in amplitudes from regional ground motion simulations, for the purpose
of validating simulations versus observations from empirical networks, and for quantifying the role of geological
features in the observed variation.
Collaboration in engineering and risk analysis
● Assess the performance of distributed infrastructure systems using simulated ground motions. Evaluate the
potential impact of basin effects, rupture directivity, spatial distribution of ground motion, or other phenomena
on risk to infrastructure systems.
● Enhance the reliability of simulations of long period ground motions in the Los Angeles region using
refinements in source characterization and seismic velocity models, and evaluate the impacts of these ground
motions on tall buildings and other long-period structures (e.g., bridges, waterfront structures).
● Identify the sensitivity of structural response to ground motion parameters and structural parameters through
end-to-end simulation. Buildings of particular interest include non-ductile concrete frame buildings.
● Perform detailed assessments of the results of ground motion simulation scenarios, as they relate to the
relationship between ground motion characteristics and structural response and damage.
● Develop improved site/facility-specific and portfolio/regional risk analysis (or loss estimation) techniques and
tools, and implement them in software tools.
● Identify earthquake source and ground motion characteristics that control damage and financial loss
estimates.
● Proposals for other innovative projects that would further implement SCEC information and techniques in
seismic hazard, earthquake engineering, risk analysis, and ultimately loss mitigation, are encouraged.
6. Research by Special Projects Working Groups
Special Projects are organized around large-scale projects funded through special grants outside of the NSF-USGS
cooperative agreements that support the SCEC base program, but have synergistic goals and are aligned with the
overall SCEC research program priorities. The current Special Projects teams include Working Group on California
Earthquake Probabilities (WGCEP), the Collaboratory for the Study of Earthquake Predictability (CSEP), the
Community Modeling Environment (CME), the Central California Seismic Project (CCSP), Collaboratory for
Interseismic Simulation and Modeling (CISM), and Mining Seismic Wavefields.
6.1 Working Group on California Earthquake Probabilities (WGCEP)
The Working Group on California Earthquake Probabilities (WGCEP) is a collaboration between SCEC, the U.S.
Geological Survey, and California Geological Survey aimed at developing official earthquake-rupture-forecast models
for California. The project is closely coordinated with the USGS National Seismic Hazard Mapping Program, and has
received financial support from the California Earthquake Authority (CEA). The WGCEP has completed the time-
independent UCERF3 model (UCERF3-TI, which relaxes segmentation and includes multi-fault ruptures) and the long-
term, time-dependent model (UCERF3-TD, which includes elastic-rebound effects). More recently, we are working to
add spatiotemporal clustering (UCERF3-ETAS) to account for the fact that triggered events can be large and damaging.
As the latter will require robust interoperability with real-time seismicity information, UCERF3-ETAS will bring us into
the realm of operational earthquake forecasting (OEF). We are also starting to plan for UCERF4, which we anticipate
will utilize physics-based simulators to a greater degree (see last bullet below).
The following research activities would contribute to WGCEP goals:
● Evaluate fault models in terms of the overall fault connectivity at depth (important for understanding the
likelihood of multi-fault ruptures) and the extent to which faults represent a well-define surface versus a proxy
for a braided deformation zone.
● Evaluate existing deformation models, or develop new ones, in terms of applicability of GPS constraints,
categorical slip-rate assignments (based on “similar” faults), applicability of back-slip methods, and other
assumptions. Of particular interest is the extent to which slip rates taper at the ends of faults and at fault
connections.
● Evaluate the UCERF3 implication that 30% to 60% of off-fault deformation is aseismic.
● Help determine the average along-strike slip distribution of large earthquakes, especially where multiple faults
are involved (e.g., is there reduced slip at fault connections?).
WELCOME
2016 SCEC Annual Meeting page 106
● Help determine the average down-dip slip distribution of large earthquakes (the ultimate source of existing
discrepancies in magnitude-area relationships). Are surface slip measurements biased with respect to slips at
depth?
● Develop a better understanding of the distribution of creeping processes and their influence on both rupture
dimension and seismogenic slip rate.
● Contribute to the compilation and interpretation of mean recurrence-interval constraints from paleoseismic
data and/or develop site-specific models for the probably of events going undetected at a paleoseismic site.
● Develop ways to constrain the spatial distribution of maximum magnitude for background seismicity (for
earthquakes occurring off of the explicitly modeled faults).
● Address the question of whether small volumes of space exhibit a Gutenberg Richter distribution of nucleations
(even on faults).
● Develop improved estimates (including uncertainties) of the total long-term rates of observed earthquakes for
different sized volumes of space.
● Refine our magnitude completeness estimates (as a function of time, space, and magnitude). Develop such
models for real-time applications (as will be needed in operational earthquake forecasting).
● Develop methods for quantifying elastic-rebound based probabilities in un-segmented fault models.
● Help quantify the amount of slip in the last event, and/or average slip over multiple events, on any major faults
in California (including variations along strike).
● Develop models for fault-to-fault rupture probabilities, especially given uncertainties in fault endpoints.
● Determine the extent to which seismicity rates vary over the course of historical and instrumental observations
(the so-called Empirical Model of previous WGCEPs), and the extent to which this is explained by aftershock
statistics.
● Determine the applicability of higher-resolution smoothed-seismicity maps for predicting the location of larger,
more damaging events.
● Explore the UCERF3 “Grand Inversion” with respect to: possible plausibility filters, relaxing the UCERF2
constraints, not over-fitting data, alternative equation-set weights, applying a characteristic-slip model, and
applicability of the Gutenberg Richter hypothesis on faults (see report at www.WGCEP.org).
● Develop applicable methods for adding spatiotemporal clustering to forecast model s(e.g., based on empirical
models such as ETAS). Are sequence-specific parameters warranted?
● Determine if there is a physical difference between a multi-fault rupture and a separate event that was
triggered quickly.
● Develop more objective ways of setting logic-tree branch weights, especially where there are either known or
unknown correlations between branches.
● Develop easily computable hazard or loss metrics that can be used to evaluate and perhaps trim logic- tree
branches.
● Develop techniques for down-sampling event sets to enable more efficient hazard and loss calculations.
● Because all models will be wrong at some level, develop valuation metrics that allow one to quantify the benefit
of potential model improvements in the context of specific uses.
● Develop novel ways of testing UCERF3, especially ones that can be integrated with CSEP. For example,
UCERF3-ETAS could be tested against historic ruptures that have occurred in the state.
● Address the extent to which large triggered events can or cannot nucleate from within the rupture area of a
main shock (the answer has an important influence on UCERF3-ETAS results).
● Help constrain the distance decay of triggered earthquakes, especially in light of a finite seismogenic thickness
and a depth-dependent rate of earthquake nucleation.
● Compile global datasets of large-event triggering, and compare with the more robust statistics available for
smaller events. Also try to quantify the extent of spatial overlap between large main shocks and large triggered
events (an important metric that could be used to tune models like UCERF3).
● Study and test the behavior of computational earthquake-cycle simulators, envisioning that they could become
essential ingredients in future UCERF projects and a cornerstone of SCEC5. The goal is to develop the
capability of simulators to be able to contribute meaningfully to hazard estimates. Important tasks include:
○ Study and test, using code verification exercises and more than one code, the sensitivity of simulator
results to input details including fault-system geometry, stress-drop values, tapering of slip, methods
of encouraging rupture jumps from fault to fault, cell size, etc.
○ Develop physically realistic ways of simulating off-fault seismicity.
WELCOME
2016 SCEC Annual Meeting page 107
○ Add additional physics into simulators, for example, the inclusion of high-speed frictional weakening
and of off-fault viscoelastic and heterogeneous elastic properties.
○ Develop alternate methods of driving fault slip besides “back-slip”.
○ Make access to existing simulators easy for new users, including adequate documentation and
version numbers, examples of input and output files for initial testing, and access to analysis tools.
Publicize availability.
○ Develop new approaches to designing simulators and/or of making them more computationally
efficient, including the use of better algorithms, point source Greens functions, and GPUs.
○ Develop validation tools for simulators, utilize existing UCERF data comparison tools with them, and
develop capabilities for simulators to interact with UCERF infrastructure.
○ Develop the capability of simulators to deal with UCERF and SCEC CFM fault geometries, both for
rectangular and triangular cell representations.
○ Create statewide synthetic earthquake catalogs spanning 100 My using as many different simulators
as possible, in order to generate statistically significant behavior on even slow-slipping faults. Use
small time-steps to permit evaluation of short-term clustering.
○ Use these catalogs as synthetic laboratories for CSEP testing as described under CSEP.
○ Data-mine these catalogs for statistically significant patterns of behavior. Evaluate whether much-
shorter observed catalogs are statistically distinguishable from simulated catalogs. Consider and
explore what revisions in simulators would make simulated catalogs indistinguishable from observed
catalogs.
○ Develop and test a variety of statistical methods for determining the predictability of earthquakes in
these simulated catalogs, especially with respect to large triggered earthquakes.
○ Compute other data types such as gravity changes, surface deformation, InSAR images, in order to
allow additional comparisons between simulated results and observations.
Further suggestions and details can be found at http://www.WGCEP.org, or by contacting the project leader (Ned Field:
[email protected]; (626) 644-6435).
6.2 Collaboratory for the Study of Earthquake Predictability (CSEP)
CSEP is developing a virtual, distributed laboratory—a collaboratory—that supports a wide range of scientific prediction
experiments in multiple regional or global natural laboratories. This earthquake system science approach seeks to
provide answers to the questions: (1) How should scientific prediction experiments be conducted and evaluated? and
(2) What is the intrinsic predictability of the earthquake rupture process?
6.2.1. Priorities for CSEP
● Retrospective Canterbury experiment. Finalize the retrospective evaluation of physics-based and statistical
forecasting models during the 2010-12 Canterbury, New Zealand, earthquake sequence by (i) comparing
retrospective forecasts against extent prospective models, (ii) transitioning models to prospective evaluation,
including in other regions.
● Global CSEP experiments. Develop and test global models, including, but not limited to, those developed
for the Global Earthquake Model (GEM).
● Strengthening testing and evaluation methods. Develop computationally efficient performance metrics of
forecasts and predictions that (i) account for aleatory variability and epistemic uncertainties, and (ii) facilitate
comparisons between a variety of probability-based and alarm-based models (including reference models).
● Supporting Operational Earthquake Forecasting (OEF). (i) Developing forecasting methods that explicitly
address real-time data deficiencies, (ii) update forecasts on an event basis and evaluating forecasts with
overlapping time-windows or on an event basis, (iii) improve short-term forecasting models, (iv) develope
prospective and retrospective experiments to evaluate OEF candidate models.
● Earthquake rupture simulators. Develop experiments to evaluate the predictive skills of earthquake rupture
simulators, against both synthetic (simulated) and observed data (see also the WGCEP section), with specific
focus on how to automate the identification of a large earthquake with a modeled fault.
● External Forecasts and Predictions (EFP). Develop and refine experiments to evaluate EFPs (generated
outside of CSEP), including operational forecasts by official agencies and prediction algorithms based on
seismic and electromagnetic data;
WELCOME
2016 SCEC Annual Meeting page 108
● Induced seismicity. Develop models and experiments to evaluate hypotheses of induced seismicity, e.g. in
the Salton Trough or in Oklahoma, including providing data access to injection/depletion rates and other
potentially pertinent data.
● Hybrid/ensemble models. Develop methods for forming optimal hybrid and ensemble models from a variety
of existing probability-based or alarm-based forecasting models;
● Hazard models. Develop experiments to evaluate seismic hazard models and their components (e.g., ground
motion models).
● Coulomb stress. Develop forecasting models based on the Coulomb stress hypothesis that can be tested
retrospectively and prospectively within CSEP.
● Focal mechanisms. Develop methodology to forecast focal mechanisms and evaluating the skill of such
forecasts.
● Paleo-based forecasts. Develop experiments to prospectively test the fault rupture and earthquake
probabilities implied by paleoseismic investigations of California faults (e.g., testing probabilities of future
ruptures at paleoseismic sites where numerous ruptures have been documented, the relative effectiveness of
proposed fault segment boundaries at stopping ruptures, and the relative frequency of on-fault and off-fault
ruptures in California) (see also the WGCEP and SoSafe sections).
6.2.2 General Contributions
● Establish rigorous procedures in controlled environments (testing centers) for registering prediction
procedures, which include the delivery and maintenance of versioned, documented code for making and
evaluating predictions including intercomparisons to evaluate prediction skills.
● Construct community-endorsed standards for testing and evaluating probability-based, alarm-based, fault-
based, and event-based predictions.
● Develop hardware facilities and software support to allow individual researchers and groups to participate in
prediction experiments.
● Design and develop programmatic interfaces that provide access to earthquake forecasts and forecast
evaluations.
● Provide prediction experiments with access to data sets and monitoring products, authorized by the agencies
that produce them, for use in calibrating and testing algorithms.
● Characterize limitations and uncertainties of such data sets (e.g., completeness magnitudes, source
parameter and other data uncertainties) with respect to their influence on experiments.
● Expand the range of physics-based models to test hypotheses that some aspects of earthquake triggering are
dominated by dynamic rather than quasi-static stress changes and that slow slip event activity can be used to
forecast large earthquakes.
● Evaluate hypotheses critical to forecasting large earthquakes, including the characteristic earthquake
hypothesis, the seismic gap hypothesis, and the maximum-magnitude hypothesis;
● Conduct workshops to facilitate international collaboratories. A major focus of CSEP is to develop international
collaborations between the regional testing centers and to accommodate a wide-ranging set of prediction
experiments involving geographically distributed fault systems in different tectonic environments.
6.3 Community Modeling Environment (CME)
The Community Modeling Environment (CME) is a SCEC special project that develops improved ground motion
forecasts by integrating physics-based earthquake simulation software, observational data, and earth structural models
using advanced computational techniques including high performance computing (HPC). CME is developing ground
motion simulations that produce broadband seismograms and often use results, and integrate work, from various SCEC
activities. The simulation tools used in CME activities include rupture generators (dynamic and kinematic), wave
propagation models (low and high frequency), non-linear site response modules, and validation capabilities (including
assembled observational strong motion data sets and waveform-matching goodness of fit algorithms and information
displays).
6.3.1 SCEC Computational Platforms
The SCEC community can contribute research activities to CME by providing scientific or computational capability that
can improve ground motion forecasts for any of the activities described below.
WELCOME
2016 SCEC Annual Meeting page 109
Broadband Platform. The open-source Broadband Platform (BBP) provides a verified, validated, and user-friendly
computational environment for generating broadband (0-100Hz) ground motions. The BBP includes a suite of kinematic
source models, wave propagation codes and site response modules to calculate suites of synthetic seismograms from
user-specified rupture sets, structural models, and station sets.
High-F Platform. The High-F platform comprises source and wave propagation codes (kinematic ruptures with the
AWP-ODC wave propagation code and Hercules for both) used by SCEC researchers to push earthquake simulations
to higher frequencies (> 1Hz). High-F activities aim at including more realistic physics while improving the upper
frequency limits of physics-based ground motions. Physics models currently tested in the High-F platform include those
for fault roughness, near-fault plasticity, frequency-dependent attenuation, topography, small-scale near-surface
heterogeneities, and near-surface nonlinearities.
CyberShake Platform. The CyberShake Platform provides physics-based probabilistic seismic hazard curves and
maps using seismic reciprocity to generate large ensembles of ground motion simulations (> 108). CyberShake
combines an earthquake rupture forecast (currently UCERF2) with a kinematic source generator (Graves and Pitarka)
and a wave propagation code (AWP-ODC) to provide rates of exceedance of ground motions.
F3DT Platform. This platform integrates the software needed for full-3D waveform tomography using the adjoint-
wavefield and scattering-integral formulations of the structural inverse problem. F3DT can invert both earthquake
waveforms and ambient-field correlograms for high-resolution crustal models, and it can refine the centroid moment
tensors of earthquakes by matching observed waveforms with 3D synthetics.
Unified Community Velocity Model Platform. The UCVM platform provides an easy-to-use software framework for
comparing and synthesizing 3D Earth models and delivering model products to users. UCVM is used as a repository
and delivery system for Community Velocity Models (CVMs) developed by SCEC researchers. UCVM allows the
generation of large simulation meshes which can be used for High-F and CyberShake simulations.
6.3.2 Research Priorities
In addition to specific priorities described in the Ground Motions section, the Planning Committee is seeking proposals
to:
● Provide codes to be included in an upcoming dynamic rupture platform, DynaShake. DynaShake is intended
to operate in a fashion similar to the BBP, where independent codes are to be integrated to run simulation
problems and go through validation exercises for realistic applications. Dynamic rupture code developers must
be willing to work under an open source license. Preference will be given to codes that have been verified
under the Dynamic Rupture Code Verification Technical Activity Group and to codes that have been
parallelized or are in the process of being parallelized to run on multiple core systems (clusters and HPC).
● Improve and integrate models for source generation, wave propagation, and site effects into any of CME’s
simulation platforms. This can include the development of scientific software that simulates one or more
physical processes from the source to the surface.
● Improve our ability to extend ground motion simulations to higher frequencies and to improve the accuracy of
such models through the integration of better physics. Proposals can be targeted to any of the ground motion
simulations platforms, but are most relevant in the context of High-F.
● Develop the computational and integration frameworks to extend 3D structural modeling capabilities to the
BBP.
● Develop or improve ground motion simulation validation computational and organizational tools. Research in
this area would contribute to the efforts under the ground motion simulation validation (GMSV) technical
activity group and the EEII. Validation of ground motions and models is important for any and all of the
simulation platforms under CME.
● Improve the accuracy of local and statewide community velocity models (CVMs), through the development of
analysis techniques. Such techniques may involve, for example, the development of 3D tomography codes
as well as the integration of geology constraints into CVMs. Proposals are also sought to improve the
methodologies used for integration of models from different sources and scales within UCVM.
● Develop tools for decimating the UCERF3 model for use within CyberShake for specific regions (Southern
and Central California) and accounting for the loss of precision and accuracy from the simplified model.
WELCOME
2016 SCEC Annual Meeting page 110
6.4 Central California Seismic Project (CCSP)
The largest uncertainties in the estimation of the catastrophic risks to California utilities come from the seismic hazard
uncertainties at low exceedance probabilities. Recent analyses indicate that these are dominated by the uncertainties
in path effects; i.e., in the prediction of strong ground motions at a fixed surface site from specified seismic sources.
SCEC has joined the Pacific Gas & Electric Company (PG&E) in developing a long-term research program aimed at
reducing the uncertainties in seismic hazard estimation with a particular emphasis of reducing the uncertainty in path
effects.
A pilot project focused on the central coast of California was initiated in 2015 and continued through 2016. The goal of
the CCSP is to assess the effectiveness of physics-based ground motion modeling in reducing path-effect uncertainties.
6.4.1 Research Strategies
Currently planned objectives of the program are fourfold:
● Analyze the existing seismic, geophysical, and geologic data for constraints on the 3D crustal structure of
Central California. The seismic constraints include earthquake waveforms and ambient-field correlograms; the
geologic constraints include surface and subsurface data on basin, fault, and basement structure.
● Invert the seismic and geologic constraints to improve models of Central California crustal structure. Priority
will be given to full-3D tomographic methods that can account for 3D wave propagation and the nonlinearity
of the structural inverse problem.
● Deploy an array of temporary seismic stations in Central California to collect new earthquake and ambient-
field data. Assess the efficacy of these data in reducing path-effect uncertainties and validating model-based
uncertainty reductions. Both ground and ocean bottom seismometer (OBS) array deployments will be
considered.
● Compute large ensembles of earthquake simulations for central California sites that are suitable for
probabilistic seismic hazard analysis (PSHA). Compare the simulation results with those from ground motion
prediction equations (GMPEs). Use this modeling to understand the aleatory variability encoded by the
GMPEs and to assess the epistemic uncertainties in the simulation-based PSHA.
6.4.2 Research Priorities
● Incorporate data from ocean bottom seismometer observations into improved community velocity models
near- and off-shore Central California.
● Improve understanding of the fault system, both onshore and offshore, in Central California using precise
earthquake locations, high-resolution geophysical imaging surveys, and other methods.
● Use observations of ground motion from local earthquakes, and dense recordings of ground motion (where
available) to characterize the ability to predict the intensity of strong ground motion and its variability.
● Improve characterization of historical earthquakes in the region, including their location, mechanism, and
finite-source characteristics (if relevant).
● Document earthquake sources and geomorphic constraints on extreme ground motions in the southern Sierra
Nevada.
● Extend geology, source characterization and wave propagation modeling work useful in ground motion
modeling to the Central California region.
In evaluating CCSP-targeted proposals, the Planning Committee will consider the relevance of the proposed work to
the overall project plan and the ability of investigators to deliver timely results during the pilot study. The PC will also
consider novel approaches to the uncertainty-reduction problem in addition to those explicitly listed in the project plan.
6.5 Collaboratory for Interseismic Simulation and Modeling (CISM)
The Collaboratory for Interseismic Simulation and Modeling (CISM) is an effort to forge physics-based models into
comprehensive earthquake forecasts using California as its primary test bed. Short-term forecasts of seismic
sequences, in combination with consistent long-term forecasts, are critical for reducing risks and enhancing
preparedness. CISM involves multiple specialties and groups, including those from WGCEP and CSEP and those
involved with earthquake simulators. The Planning Committee is welcoming proposals that support CISM in its mission
to improve predictability by combining rupture simulators that account for the physics of rupture nucleation and stress
transfer with ground-motion simulators that account for wave excitation and propagation. CISM forecasting models will
be tested against observed earthquake behaviors within the existing CSEP. Detailed research priorities associated with
WELCOME
2016 SCEC Annual Meeting page 111
CISM include the use and calibration of earthquake simulators and priorities listed under the WGCEP and CSEP
sections.
6.6 Mining Seismic Wavefields (MSW)
The Mining Seismic Wavefields (MSW) project was initiated in 2016 to develop and deploy cyberinfrastructure for
mining seismic wavefields through data intensive computing techniques to extend similarity search for earthquake
detection to massive data sets. Similarity search has been used to understand the mechanics of tectonic tremor,
transform our understanding of the depth dependence of faulting, illuminate diffusion within aftershock seismicity, and
reveal new insights into induced earthquakes. These results were achieved with modest data volumes – from ~ 10
seismic stations spanning ~ 10km – yet they increased the number of detected earthquakes by a factor of 10 to 100.
This project will develop the cyberinfrastructure required to enable high-sensitivity studies of earthquake processes
through the discovery of previously undetected seismic events within massive data volumes. The Planning Committee
welcomes proposals that support the Mining Seismic Wavefields project. Key priorities can be found in the Seismology
section. In addition, we encourage proposals to develop and apply data-intensive computing techniques, e.g. for long
seismic time series (large-T) or dense seismic instrumentation (large-N) analysis, to earthquake science and
engineering research problems.
7. Communication, Education, and Outreach Activities
SCEC’s Communication, Education, and Outreach (CEO) program facilitates learning, teaching, and application of
earthquake research. In addition, SCEC/CEO has a global public safety role in line with the third element of SCEC’s
mission: “Communicate understanding of earthquake phenomena to end-users and society at large as useful
knowledge for reducing earthquake risk and improving community resilience.” The theme of the CEO program in
SCEC5 is Partner Globally, Prepare Locally. Our geographic reach has been expanded far beyond the Golden State.
Our partnerships foster new research opportunities and ensure the delivery of research and educational products that
improve the preparedness of the general public, government agencies, businesses, research and practicing engineers,
educators, students, and the media—locally in California as well as in other states and countries. Prepare Locally not
only refers to improved resiliency to local hazards, but also to preparing students and the public for the future with the
enhanced science literacy to make informed decisions to reduce their risk, and to preparing future scientists via
research opportunities and support through career transitions. In SCEC5, the CEO program will manage and expand
activities within four CEO focus areas:
1. Knowledge Implementation connects SCEC scientists and research results with practicing engineers,
government officials, business risk managers, and other professionals active in the application of earthquake
science.
2. The Public Education and Preparedness focus area will educate people of all ages about earthquakes,
tsunamis, and other hazards, and motivate them to become prepared.
3. The K-14 Earthquake Education Initiative will improve Earth science education in multiple learning
environments, overall science literacy, and earthquake safety in schools and museums.
4. Finally, the Experiential Learning and Career Advancement program will provide research opportunities,
networking, and other resources to encourage and sustain careers in STEM fields.
Investigators interested in contributing to CEO activities are strongly advised to contact Mark Benthien (213-740-0323;
[email protected]) before submitting a proposal since alternative approaches may be more appropriate.
WELCOME
2016 SCEC Annual Meeting page 112
Saturday, September 10
08:00 - 09:00 Workshop Check-In at Plaza Ballroom Foyer of Hilton Palm Springs
09:00 - 17:00 SCEC SoSAFE Workshop: Recent Successes and Future Challenges
Conveners: Kate Scharer (USGS) and Ramon Arrowsmith (ASU) Location: Palm Canyon Room
09:00 Introductions
Earthquake Recurrence and Slip Over Short and Long Term: How Does It All Add Up?
Creep and other earthquake cycle challenges from recent observations
09:15 The Dry Lake Valley site: Observations of structures formed by modern and prehistoric creep on the central San Andreas fault (Nate Toke)
09:30 South Napa 2013 trenches: creep, afterslip, missed events! (Tim Dawson)
09:45 Creep on the Imperial Fault and new faults in the Salton Trough (Eric Lindsey)
10:00 Near field and off fault deformation revealed using optical image correlation (Chris Milliner)
10:15 Break
Adding up slip: Segment boundaries and slip rate variations
10:30 Slip rates and distributed deformation in and around San Gorgonio Pass (Michele Cooke)
10:45 Earthquake displacements and timing of events at Quincy and Mystic Lake, SJF (Nate Onderdonk)
11:00 Investigating the age and origin of small offsets at Van Matre Ranch along the San Andreas Fault in the Carrizo Plain, California (Barrett Salisbury)
11:15 Discussion
- What does slip rate variability mean for fault and crustal rheologies?
- Why aren't paleoseismic/geologic data used as constraints on this?
- Proposal and project ideas
12:00 Lunch
Integrating Earthquake and Paleoclimate/Paleoenvironmental Chronologies on the SoSAFE System
13:00 A lake-based event chronology for the Southern San Andreas and San Jacinto Faults (Tom Rockwell)
13:15 Sediment accumulation curves discriminate between proximal event records on the southern San Andreas Fault (Kate Scharer)
13:30 Holocene droughts, fires, floods and pluvials in southwestern California (Matt Kirby)
14:00 On the PAGES2k project, paleoclimate cyberinfrastructure and integrating paleoclimatology and paleoseismology Nick McKay
14:30 Discussion
- Where does SoSAFE go to advance this research?
15:00 Break
WELCOME
2016 SCEC Annual Meeting page 113
Saturday, September 10 Outside Looking In: Broad Applications of Behavior of High Slip Rate Faults
15:15 So many earthquakes, so little time: An examination of SAF-system earthquake recurrence (Glenn Biasi)
15:30 RSQSim for paleoseismologists; what’s under the hood and SoCal results (Jacqui Gilchrist)
16:00 Data integration and visualization tools for bringing paleoseismic data and simulator results together (Kevin Milner)
16:15 Discussion and Wrap Up
- Earthquake geology data and metadata needs
- Paleoseismology in the Collaboratory for Interseismic Simulation and Modeling (CISM)
- Simulator opportunities
- Adieu to SoSAFE!
16:45 Adjourn
09:00 - 17:00 SCEC Ventura Special Fault Study Area Workshop
Conveners: Scott Marshall (AppState), James Dolan (USC), Thomas Rockwell (SDSU), and John Shaw (Harvard)
Location: Plaza Ballroom AB
09:00 Opening Remarks and Introduction to the Ventura fault system (James Dolan, Tom Rockwell, John Shaw)
Crustal Structure, Near Surface Geophysics, Seismology
09:15 Subsurface Structure of the Western Transverse Ranges and Potential for Large Earthquakes (Yuval Levy)
09:30 3D Fault Geometry and Slip History of the North Channel-Pitas Point-Red Mountain Fault System (Craig Nicholson)
09:45 Constraints on the Geometry of the Ventura-Pitas Point Fault System at Depth & Implications for Hazards (John Shaw)
10:00 Moderated Discussion
10:30 Break
Geology, Paleoseismology, Tectonic Geomorphology
10:45 Evidence for Large (~M8.0) Multi-Segment Ruptures in the Ventura Pitas Point Fault System (Tom Rockwell)
11:00 Urban Paleoseismology of the Ventura Fault and Evidence for Large Coseismic Uplift Events (James Dolan)
11:15 Geomorphic Evidence for Recent Deformation along the Southern San Cayetano Fault (Dylan Rood)
11:30 Stress Drop Variations and Seismicity Characteristics in San Gorgonio and Ventura (Thomas Goebel)
11:45 Moderated Discussion
12:15 Lunch
Geodesy and Fault Modeling
13:00 Regional Geodetic Fault Slip Rates and Interseismic Deformation (Scott Marshall)
13:15 Uplift Across the Western Transverse Ranges from Integrated Analysis of GPS, InSAR, Leveling and Tide Gauge Data (Bill Hammond)
13:30 Kinematic Models of Fault Slip Rates in the Western Transverse Ranges (Kaj Johnson)
13:45 Moderated Discussion
14:15 Break
WELCOME
2016 SCEC Annual Meeting page 114
Saturday, September 10 Dynamic Rupture and Tsunamis
14:30 Searching for Tsunami Deposits in the Santa Barbara Channel (Alex Simms)
14:45 Dynamic Rupture and Tsunami Models Offshore Ventura (Kenny Ryan)
15:00 Tsunamis from Earthquakes on the Ventura-Pitas Point Fault System (Hong Kie Thio)
15:15 Moderated Discussion
15:45 Break
What's Next? Future Research Directions
16:00 Overview of Final Report to SCEC
16:15 Final Discussion: Plans for Future Work and/or Special Publication
16:45 Final Remarks (Scott Marshall)
17:00 Adjourn
09:00 - 17:00 SCEC on the Processes that Control the Strength of Faults and Dynamics of Earthquakes
Conveners: David Goldsby (U Penn), Whitney Behr (UT Austin), Eric Dunham (Stanford) and Greg Hirth (Brown)
Location: Plaza Ballroom CD
09:00 Workshop Overview and Introductions (David Goldsby/Whitney Behr)
Mechanisms of coseismic weakening
09:10 Controls on frictional strength and stability in gouge-filled faults in the seismogenic zone (Chris Spiers)
09:35 The role superplasticity may play during seismic slip (Nicola De Paola)
10:00 How coseismic weakening can be incorporated into earthquake models and linked with seismic data (Rob Viesca)
10:25 Group Discussion
10:55 Break
Spatial Variations in fault resistance
11:10 Field measurements of fault roughness (Emily Brodsky)
11:35 How roughness influences the dynamics of an earthquake in numerical models (Eric Dunham)
12:00 Lunch
13:00 How changes in coseismic weakening within the seismogenic zone limit the depth penetration of large earthquakes (Nadia Lapusta)
13:25 Earthquake source scaling (Marine Denolle)
13:50 Group Discussion
14:20 Break
Evolution of the slip zone during an earthquake
14:35 Recent progress identifying field evidence for seismic slip (Heather Savage)
15:00 How dynamic stresses influence the properties of fault damage fault zones (Ashley Griffith)
15:25 How chemical alteration of slip surfaces can be linked to coseismic weakening mechanisms such as flash heating (Jim Evans)
15:50 Group Discussion
Where we are and whither we are tending
16:20 Summary discussion / Wrap up
17:20 Adjourn
WELCOME
2016 SCEC Annual Meeting page 115
Sunday, September 11 07:00 - 17:00 SCEC Annual Meeting Registration & Check-In at Hilton Lobby
07:00 - 08:00 Breakfast at Hilton Poolside
08:00 - 12:00 SCEC Collaboratory for Interseismic Simulation and Modeling (CISM) Workshop
Convener: Thomas Jordan (USC) Location: Horizon Ballroom
07:30 Continental Breakfast & Check-In
08:00 Introduction to CISM (Tom Jordan)
08:20 Multiscale forecasting using UCERF3 (Ned Field)
09:00 Earthquake forecasting using the RSQSim earthquake simulator (Jim Dieterich)
09:40 Contributions of the UseIT summer research program to CISM (Kevin Milner)
10:00 Generating long RSQSim catalogs for CISM analysis (Jacqui Gilchrist)
10:00 Discussion: Use of simulators in CISM forecasting research (All)
10:15 Break
10:30 Testing forecasting models in CSEP (Max Werner)
11:00 Is there an earthquake drought in California? (Dave Jackson)
11:30 CyberShake as a CISM ground motion prediction platform (Scott Callaghan)
11:45 Discussion: CISM plans for 2017 (All)
12:00 Adjourn and Lunch (Hilton Restaurant and Tapestry Room)
13:00 - 17:00 International Workshop on Ground Motion Simulation Validation
Conveners: Sanaz Rezaeian (USGS), Nicolas Luco (USGS), Brendon Bradley (QuakeCoRE), Iunio Iervolino (U Naples), and Leonardo Ramirez-Guzman (UNAM)
Location: Horizon Ballroom
13:00 Introductions (Sanaz Rezaeian)
13:05 The SCEC perspective on validations in SCEC5 (Christine Goulet)
Overview of GMSV Efforts in New Zealand, Italy, Mexico, and at SCEC
13:15 QuakeCoRE GMSV research coordination and current priorities (Brendon Bradley)
13:25 Italian experience with 3D physics-based numerical simulations of earthquake ground motion (Marco Stupazzini)
13:40 Synthetic ground motion system for structural design and evaluation in Mexico City (Leonardo Ramirez-Guzman)
13:55 Organization and structure of the GMSV TAG in SCEC4 (Nico Luco)
14:05 Discussion: How should modelers and engineers participate in and contribute to the SCEC5 ground motion simulation validation efforts?
Simulated Earthquake Ground Motions and Database Development
14:30 Development of a QuakeCoRE database for access to ground motion simulation outputs at specific locations (Sung Bae)
14:40 Datasets of recorded and simulated near-source earthquake ground motions: the experience gained in Italy (Marco Stupazzini)
14:50 Current SCEC simulations (Phil Maechling)
15:00 Discussion: What are the requirements of a database interface for users to access simulated ground motions and validation parameters?
WELCOME
2016 SCEC Annual Meeting page 116
Sunday, September 11
15:25 Break
Validation Efforts and Engagement with Engineering Users
15:40 QuakeCoRE guidance on the utilization of ground motion simulations in engineering practice (Brendon Bradley / Jack Baker)
15:50 Engineering applications of 3D physics-based numerical simulations: two case studies from the Reluis project, Italy (Marco Stupazzini)
16:00 Demonstration of the efficacy of the BBP validation gauntlets for building response analysis applications (Nico Luco / Greg Deierlein)
16:15 Discussion: How should we engage engineering users in SCEC5?
16:45 Summary and Wrap Up (Sanaz Rezaeian)
17:00 Adjourn
13:00 - 17:00 Workshop on Science Communication: Navigating and Maximizing a Digital, Social World
Conveners: Wendy Bohon (IRIS), Beth Bartel (UNAVCO, and Jason Ballmann (SCEC) Location: Palm Canyon Room SCEC Communication, Education, and Outreach is pleased to announce an afternoon of social
media discussions and media training at the SCEC Annual Meeting on Sunday, September 11th, between 1 PM and 5 PM, featuring award-winning earthquake journalist for the LA Times, Ron Lin, and communications experts and scientists from IRIS, UNAVCO, and SCEC.
Join with us in a casual, interactive environment where you will leave knowing best practices for social media networking, how to best prepare for and talk to the media, and where it’s all going next. You might start a Twitter account while there, discover ways to follow your colleagues online, or meet a journalist or two who can bring your research to the right audience. There is no need to RSVP for these sessions. Refreshments will be served.
(1) Improving Your Online Presence: Social Media, Web Profiles, Emails, Blogs: Dr. Wendy
Bohon (IRIS), Beth Bartel (UNAVCO), and Jason Ballmann (SCEC) will lead a panel, Q&A, and mini-trainings on how to outfit your digital prominence, with the goals of networking with other scientists and experts, best practices on sharing information publicly, and how to engage certain mediums like social media, blogs, emails, and websites.
(2) Talking to the Media, with Award-Winning LA Times reporter Ron Lin Learn about the story-
writing process, what journalists look for and need to make a compelling article, prime examples of successful earthquake science articles, and how you can look for ways to make those broader impacts with your science!
13:00 Introduction
13:30 Key Online Mediums You Need to Know About
14:00 Who to Follow and Why (Following Spree!)
14:30 Sharing Information Publicly, Promoting Your Research
15:30 Talking to the Media, with Award-Winning LA Times reporter Ron Lin
WELCOME
2016 SCEC Annual Meeting page 117
Sunday, September 11 15:00 - 20:00 Poster Set-Up in Plaza Ballroom
17:00 - 18:00 Welcome Social in Hilton Lobby and Plaza Ballroom
18:00 - 19:00 Distinguished Speaker Presentation in Horizon Ballroom
Earthquakes on Compressional Inversion Structures - Problems in Mechanics and in Hazard Assessment (Richard H. Sibson)
19:00 - 20:30 Welcome Dinner at Hilton Poolside
19:00 - 21:00 SCEC Advisory Council Meeting in Palm Canyon Room
21:00 - 22:30 Poster Session 1 in Plaza Ballroom
Monday, September 12 07:00 - 08:00 SCEC Annual Meeting Registration & Check-In at Hilton Lobby
10:00 - 15:00 SCEC Annual Meeting Registration & Check-In at Hilton Lobby
07:00 - 08:00 Breakfast at Hilton Poolside
08:00 - 10:00 Session 1: The State of SCEC in Horizon Ballroom
08:00 Welcome and State of the Center (Tom Jordan)
08:30 Agency Reports
- National Science Foundation (Greg Anderson/Carol Frost)
- U.S. Geological Survey (Bill Leith)
09:00 Communication, Education, & Outreach (Mark Benthien)
09:20 SCEC Science Accomplishments (Greg Beroza)
10:00 - 10:30 Break
10:30 - 12:30 Session 2: Modeling Fault Systems – Supercycles
Moderators: Mike Oskin, Kate Scharer
10:30 Open Intervals, Clusters and Supercycles: 1100 years of Moment Release in the Southern San Andreas Fault System: Are we Ready for the Century of Earthquakes? (Tom Rockwell)
11:00 The bridge from earthquake geology to earthquake seismology (Dave Jackson)
12:00 Discussion
12:30 - 14:00 Lunch at Hilton Restaurant, Tapestry Room, and Poolside
14:00 - 16:00 Session 3: Modeling Fault Systems – Community Models in Horizon Ballroom
Moderators: Brad Aagaard, Michele Cooke
14:00 How Sensitive are Inferred Stresses and Stressing Rates to Rheology? Clues from Southern California Deformation Models (Liz Hearn)
14:30 How stressed are we really? Harnessing community models to characterize the crustal stress field in Southern California (Karen Luttrell)
15:00 Discussion
16:00 – 17:30 Poster Session 2 in Plaza Ballroom
19:00 - 21:00 SCEC Honors Banquet at Woodstock Ballroom, Hard Rock Hotel
21:00 - 22:30 Poster Session 3 in Plaza Ballroom
WELCOME
2016 SCEC Annual Meeting page 118
Tuesday, September 13
07:00 - 08:00 Breakfast at Hilton Poolside
08:00 - 10:00 Session 4: Understanding Earthquake Processes in Horizon Ballroom; Moderators: Nick Beeler,
Nadia Lapusta
08:00 Constraints on the Source Parameters of Low-Frequency Earthquakes in Parkfield and Cascadia (Amanda Thomas)
08:30 Kumamoto earthquake: a complex earthquake sequence with large strike-slip ruptures (Koji Okumura)
09:00 Discussion
10:00 - 10:30 Break
10:30 - 12:30 Session 5: New Observations in Horizon Ballroom;
Moderators: Yehuda Ben-Zion, Gareth Funning
10:30 The Ups and Downs of Southern California: Mountain Building, Sea Level Rise, and Earthquake Potential from Geodetic Imaging of Vertical Crustal Motion (Bill Hammond)
11:00 Offshore Pacific-North America lithospheric structure and Tohoku tsunami observations from a southern California ocean bottom seismometer experiment (Monica Kohler)
11:30 Discussion
12:30 - 14:00 Lunch at Hilton Restaurant, Tapestry Room, and Poolside
14:00 - 16:00 Session 6: Characterizing Seismic Hazard - Operational Earthquake Forecasting in Horizon
Ballroom Moderators: Ned Field, Max Werner
14:00 Induced earthquake magnitudes are as larger as (statistically) expected (Nick van der Elst)
14:30 Blurring the boundary between earthquake forecasting and seismic hazard (Matt Gerstenberger)
15:00 Discussion
16:00 - 17:30 Poster Session 4 in Plaza Ballroom
19:00 - 21:00 Dinner at Hilton Poolside
19:00 - 21:00 SCEC Advisory Council Meeting in Boardroom
21:00 - 22:30 Poster Session 5 in Plaza Ballroom
22:30 - 23:00 Poster Removal from Plaza Ballroom
WELCOME
2016 SCEC Annual Meeting page 119
Wednesday, September 14
07:00 - 08:00 Breakfast at Poolside
08:00 - 08:30 Report from the Advisory Council (John Vidale)
08:30 - 10:30 Session 7: Reducing Seismic Risk in Horizon Ballroom
Moderators: Jack Baker, Christine Goulet
08:30 Utilization of earthquake ground motions for nonlinear analysis and design of tall buildings (Greg Deierlein)
09:00 Progress Report of the SCEC Utilization of Ground Motion Simulations (UGMS) Committee (C.B. Crouse)
09:30 Discussion
10:30 - 12:00 The Future of SCEC in Horizon Ballroom
10:30 This Next Year: 2017 SCEC Science Collaboration (Greg Beroza)
11:30 Towards SCEC5 Priorities (Tom Jordan)
12:00 Adjourn
12:30 - 14:30 SCEC Planning Committee Lunch Meeting in Palm Canyon Room
12:30 - 14:30 SCEC Board of Directors Lunch Meeting in Tapestry Room
WELCOME
2016 SCEC Annual Meeting page 120
Distinguished Speaker Presentation Sunday
Earthquakes on Compressional Inversion Structures - Problems in Mechanics and in Hazard Assessment,
Richard H. Sibson (University of Otago)
Sunday, September 11, 2016 (18:00)
The dip-distribution of near-pure reverse-slip ruptures includes a dominant ‘Andersonian’ cluster (dips of 30±5°) flanked by
groups of low-angle thrusts (dips of 10±5°), and moderate-steep reverse faults (dips of 50±5°). These last are attributable in part
to ‘compressional inversion’ – reactivation during crustal contraction of normal faults inherited from earlier crustal extension –
identified by a distinctive structural-stratigraphic signature and shown by seismic profiling to be widespread. It is easier, in terms
of required stress levels, to impose brittle deformation on the crust during extension; compressional inversion is thus an inevitable
consequence of the Wilson Cycle of oceanic opening and closure. Examples of seismically active inversion provinces include NE
Honshu, Japan, and the NW South Island, NZ, with damaging inland earthquakes < M7.5.
Reactivation mechanics does much to explain the observed dip distribution for reverse-slip ruptures, suggesting first, that low-
displacement faults are characterized by ‘Byerlee’ friction (μs ~ 0.6), and second, that high fluid overpressures are needed for
continued reactivation of moderate-steep reverse slip faults. Support for the latter comes from the existence of hydrothermal
vein systems formed by ‘fault-valve’ action around reverse faults exhumed from seismogenic depths, and from geophysical
anomalies (bright-spot reflectors, anomalously high Vp/Vs, high electrical conductivity) in the mid-crust of areas undergoing
inversion. Compressional inversion earthquakes are predominantly ‘fluid-driven’ (H2O, CO2?) because failure on such structures
is likely induced by accumulating fluid overpressure rather than by increasing differential stress.
Evaluating hazard from inversion structures is problematic because their surface expression is often structurally complex with
misleading dip-separations, and may be obscured near the margins of sedimentary basins. Complexity also arises from
competition between inversion structures and younger, more optimally oriented thrusts. Inversion structures are likely widespread
throughout North America within the west coast plate boundary zone (e.g. Ventura and Los Angeles Basins within the Transverse
Ranges), within mid-continental rift zones, and along the Atlantic seaboard where inherited Mesozoic rift structures are
occasionally reactivated in reverse-slip earthquakes (e.g. 1982 Miramachi, New Brunswick; 1983 Goodnow, NY).
Richard (Rick) Sibson graduated BScHons (1st Class) in Geology from the University of Auckland in 1968 before
gaining an MSc and PhD (1977) from Imperial College, London, researching the structure of the Outer Hebrides
Thrust zone. He taught structural geology at Imperial College (1973-1981) and at UC Santa Barbara (1982-
1990), before returning to New Zealand as Professor of Geology in the University of Otago (1990-2009). His
research focuses on the structure of crustal fault zones and the mechanics of shallow crustal earthquakes with
coupled interests in crustal fluid flow and structural controls affecting the development of fault-hosted mineralisation. He has
been a contributing author to over 95 research papers and has presented short courses to the mineral industry and academia in
Australia, Canada, the United States, South Africa, Chile, Japan, Taiwan, and Italy.
He was elected to fellowship of the Geological Society of America (1991); the Royal Society of New Zealand (1993); the
American Geophysical Union (1999); the Royal Society of London (2003); the American Association for the Advancement of
Science (2005); and the Society of Economic Geologists (2010). Recent awards include the Wollaston Medal from the Geological
Society of London (2010), a Career Contribution Award from GSA Structural Geology and Tectonics Division (2011), and
appointment as the 2012 AGU Francis Birch lecturer, presenting “Inside a Crustal Earthquake: the Rock Evidence”.
WELCOME
2016 SCEC Annual Meeting page 121
Plenary Talk Presentations Monday
Open Intervals, Clusters and Supercycles: 1100 years of Moment Release in the Southern San Andreas
Fault System: Are we Ready for the Century of Earthquakes? Thomas K. Rockwell (SDSU)
Monday, September 12, 2016 (10:30)
Compilation of paleoseismic data from several dozen trench sites in the southern San Andreas fault system, along with
geomorphic observations of displacement in recent earthquakes, allows for sequencing of the past 1100 years of large
earthquakes for the southern 160 km of the main plate boundary. At least four generalizations are clear: 1) M7 and larger
earthquakes account for most of the moment release in the southern San Andreas fault system over the past 1100 years; 2) large
earthquakes on individual faults are quasi-periodic but display a relatively high coefficient of variation in recurrence time, similar
to most long California paleoseismic records, and which may be a reflection of Coulomb stress interactions; 3) moment release
has temporally varied during the past 1100 years but within potentially predictable bounds; and 4) the southern San Andreas
fault system is currently moment deficit when compared to the previous millennium of moment release. Comparison of the
timing of earthquakes with the timing of past high-stands of Lake Cahuilla supports the idea that lake fillings influenced and
may have modulated the timing of southernmost San Andreas fault earthquakes. If so, then the long (300 year) open interval on
the southern San Andreas fault may be in part explained by the absence of a complete lake filling episode in the past 300 years.
Together, the record suggests that the southern San Andreas fault is late in the cycle but not necessarily “overdue”, and that a
systems level approach may be more accurate in long-term earthquake forecasting than estimates made from individual faults.
If the past is the key to the future, then the next century is likely to see far more large earthquake activity on the southern San
Andreas fault system than was observed in the past one to two centuries.
The bridge from earthquake geology to earthquake seismology, David D. Jackson (UCLA)
Monday, September 12, 2016 (11:00)
Recurrence intervals of large earthquakes exceed the instrumental earthquake record, so we rely on paleo-seismic and surface
deformation evidence to infer long-term earthquake rates and recurrence statistics. Paleo-seismic studies generally imply quasi-
periodic recurrence and greater rates for large earthquakes than those deduced from instrumental data coupled with Gutenberg-
Richter magnitude relations.
Under SCEC leadership a team of paleo-seismic experts compiled the most reliable data available for use in the third Uniform
California Earthquake Rupture Forecast (UCERF3). They reported dates of observed displacements at 32 sites on 13 named faults
in California. Corrected for multiple-site ruptures, the total reported paleo-event rate at those sites exceeds about 4 per century,
yet none has occurred since 1916. Such a long hiatus is extremely unlikely for a Poisson process and even less probable for an
ensemble of quasi-period processes.
Hypotheses for the hiatus of paleo-events include (1) extreme luck, (2) unexplained regional fault interaction, or (3) mistaken
identification of non-seismic displacements as evidence of large earthquakes, or counting multiple branches of single ruptures
as separate ruptures. The first can be rejected with 99% confidence. There is no evidence for the second in the pre-1916 paleo-
seismic history or in any theoretical models yet reported. Temporal clustering of large quakes (“supercycles”?) could in principle
explain the hiatus, but individual site records and rate-state frictional models suggest quasi-periodic recurrence instead. The third
hypothesis could explain the observed quiescence because mistaken identity would be prevented by instrumental seismic data
during the last century. In any case the paleo-event hiatus poses a serious challenge to earthquake forecast models. It also begs
the larger question of how 2-dimensional fault-based observations and models can be reconciled with 3-D instrumental
earthquake observations.
WELCOME
2016 SCEC Annual Meeting page 122
Plenary Talk Presentations Monday
How Sensitive are Inferred Stresses and Stressing Rates to Rheology? Clues from Southern California
Deformation Models, Elizabeth Hearn (Capstone Geophysics)
Monday, September 12, 2016 (14:00)
Variations in fault zone strength and lower lithosphere viscoelastic rheology have been shown to influence deformation patterns
in fault systems over a range of time scales (e.g. Bird and Kong, 1994; Fay and Humphreys, 2005). Because of our mandate to
characterize stresses in the southern California lithosphere, SCEC has proposed a resource (the Community Rheology Model) that
will facilitate representing Southern California’s rheology in deformation models. I will present some recent modeling results
demonstrating how assumed fault zone and rock rheologies influence estimates of stresses, stressing rates and fault slip rates.
My kinematic and dynamic finite-element models incorporate the UCERF3 block model-bounding fault geometry (Field et al.,
2014), with refinements suggested by geologists involved in the UCERF3 effort. The kinematic models are fit to either the SCEC
CMM4 velocity field or a strain energy minimization criterion by varying fault slip rates, within the UCERF3 ranges. The
(preliminary) dynamic models are tuned to fit geological fault slip rates and SHmax orientations by varying fault zone and rock
rheologies. Deformation is driven from the sides in both sets of models.
The kinematic models suggest that 22-35% of the strain accumulation across Southern California is not associated with
interseismic locking of the modeled faults, consistent with other studies (e.g. Bird, 2009, Johnson, 2013; Zheng and Shen, 2016).
This is true whether or not the GPS velocity field is corrected for possible seismic cycle perturbations due to large SAF earthquakes.
Plasticity exerts a profound influence on slip rates, deformation patterns and stresses obtained from dynamic deformation models.
Variation in fault zone strength (quantified as shear traction) also affects stress magnitudes and orientations, but not surface
velocities or slip rates in models with just the San Andreas Fault zone represented. Simulations are underway to assess how
variations in fault zone strength and lithosphere rheology affect deformation in dynamic models incorporating the full fault
network.
How stressed are we really? Harnessing community models to characterize the crustal stress field in
Southern California, Karen M. Luttrell (LSU)
Monday, September 12, 2016 (14:30)
The in situ crustal stress field fundamentally governs, and is affected by, the active tectonic processes of plate boundary regions,
yet questions remain about the characteristics of this field and the implications for active faults in the upper crust. We investigate
the nature of this stress field in southern California by combining observations from the SCEC Community Stress Model, including
stress orientation, stress from topography, and stress accumulation rate on major locked faults. First, we estimate the magnitude
of the non-lithostatic in situ stress field in southern California by balancing in situ orientation indicated by earthquake focal
mechanisms against the stress imposed by topography, which tends to resist the motion of strike-slip faults. Our results indicate
that most regions require in situ differential stress of at least 20 MPa at seismogenic depth. In the areas of most rugged
topography along the San Andreas Fault System, differential stress at seismogenic depth must exceed 62 MPa consistent with
differential stress estimates from complimentary methods. Second, we assess the origin of the heterogeneity observed in the
stress field by combining stress accumulation on major locked fault segments with stress from topography and a simple 2-D
tectonic driving stress. Our results suggest that in situ stress heterogeneity at the regional scale is more influenced by deep
driving processes acting on a laterally heterogeneous crust than by perturbations to the stress field associated with major locked
faults in the upper crust. Finally, we discuss some potential avenues for moving toward a 4D representation of the crustal stress
field in southern California.
WELCOME
2016 SCEC Annual Meeting page 123
Plenary Talk Presentations Tuesday
Constraints on the Source Parameters of Low-Frequency Earthquakes in Parkfield and Cascadia, Amanda
Thomas (U Oregon), Gregory C. Beroza (Stanford), David R. Shelly (USGS), Michael Bostock (UBC), Allan M. Rubin
(Princeton), Genevieve Savard (UBC), & Lindsey Chuang (UBC)
Tuesday, September 13, 2016 (08:00)
Low-frequency earthquakes (LFEs) are small repeating earthquakes that occur in conjunction with deep slow slip. Like typical
earthquakes LFEs are thought to represent shear slip on crustal faults but when compared to earthquakes of the same magnitude
LFEs have lower corner frequencies, implying longer durations, and are depleted in high-frequency content relative to earthquakes
of similar magnitude. Along the San Andreas Fault, and in the Cascadia subduction zone, LFEs occur in rapid succession, forming
tectonic tremor. Here we present results from LFE source studies in both Parkfield and Cascadia. In Parkfield, we use an empirical
Green’s function approach to investigate which physical properties of the LFE source cause high-frequency depletion. We find
that the M~1 LFEs have typical durations of ~0.2 s. Using the annual slip rate of the deep SAF and the average number of LFEs
per year we estimate average LFE slip rates of ~0.24 mm/s. When combined with the LFE magnitude this number implies a stress
drop of ~104 Pa, two to three orders of magnitude lower than ordinary earthquakes, and a rupture velocity of 0.35 km/s. Typical
earthquakes are thought to have rupture velocities of ~80-90% the shear wave speed. In Cascadia, we correct LFE waveforms for
path effects and use the resulting source time functions to calculate LFE duration and magnitude. We use these estimates to
show that, like Parkfield, LFEs in Cascadia also have low stress drops, rupture and slip velocities. We also find that LFE duration
displays a weaker dependence upon moment than expected for self-similarity, suggesting that LFE asperities are limited in
dimension and that moment variation is dominated by slip. This behavior implies that LFEs exhibit a scaling distinct from both
large-scale slow earthquakes and regular seismicity. Together the slow rupture velocity, low stress drops, and slow slip velocity
explain why LFEs are depleted in high frequency content relative to ordinary earthquakes and suggest that LFE asperities represent
areas capable of relatively higher slip speed in deep fault zones. Additionally, changes in rheology may not be required to explain
both LFEs and slow slip; the same process that governs the slip speed during slow earthquakes may also limit the rupture velocity
of LFEs.
Kumamoto earthquake: a complex earthquake sequence with large strike-slip ruptures, Koji Okumura
Tuesday, September 13, 2016 (08:30)
The Mw 7.3 mainshock of the Kumamoto earthquake on April 16 ruptured the Futagawa fault zone mostly as it was forecasted.
However, the sequence of earthquakes, ruptures, and localized severe shaking were beyond the forecast demonstrating the
complex nature.
The earthquake sequence began with a Mw 6.2 at 21:26 JST on April 14. This one ruptured deeper part of the Takano and
Shirahata faults (TSF) in SW of the Futagawa fault. Strong shaking (JMA scale 7 out of 7) damaged many houses in a small area
of Mashiki town-center to kill 9 people. Following another Mw 6.0 on the April 14 source area, the Mw 7.3 occurred at 01:25 JST
on April 16. The mainshock induced several Mw 5.3 to Mw 5.9 earthquakes outside its source area, at 75 km NW, 45 km NW,
and 40 km SW away from the epicenter. Those induced earthquakes and aftershocks occurred in a zone about 120 km long
across Kyushu Island while the Mw 7.3 rupture was only 30 km long.
The 30 km long Mw 7.3 rupture mostly followed the previously mapped Futagawa fault zone with dextral offset up to 2 m.
However, the surface offset of the TSF was only 0.3 m. Also, the TSF ruptured twice both on April 14 and on April 16. Assuming
the two shocks ruptured the same fault plane, the small offset on April 16 could be due to the strain release on April 14, or to
the rather short elapsed time of 1200 to 1500 years. The northeasternmost 5 km section, which also caused severe damages,
cutting into the Aso caldera was not mapped before. Rapid erosion in this high-relief area probably erased or buried past surface
WELCOME
2016 SCEC Annual Meeting page 124
ruptures. A 3.7 km long branch fault with ~1 m offset appeared from a restraining bend of the Futagawa fault to the Mashiki
town-center with conjugate sinistral offset. These faults ruptured active alluvial plain.
The severest damages were concentrated in a small area of the Mashiki town-center where JMA intensity 7 was recorded twice
during Mw 6.2 and Mw 7.0. The possible causes of the severe damages are the source process, site amplification, surface faulting,
slope failure and lateral spreading, and so on. It is necessary to integrate analyses and models by different disciplines. From
geologic points of view, the presumable complexity of shallow subsurface structures at the boundary between Holocene graben
fill in south and upland-forming Pleistocene volcanics and gravels in north is a likely cause of the localized amplification. Further
investigation is necessary to solve this significant hazard.
WELCOME
2016 SCEC Annual Meeting page 125
Plenary Talk Presentations Tuesday
The Ups and Downs of Southern California: Mountain Building, Sea Level Rise, and Earthquake Potential
from Geodetic Imaging of Vertical Crustal Motion, William C. Hammond (UNR), Geoffrey Blewitt (UNR), Corné
W. Kreemer (UNR), Reed J. Burgette (NMSU), Kaj M. Johnson (Indiana, Charles M. Meertens (UNAVCO), & Frances
Boler (UNAVCO)
Tuesday, September 13, 2016 (10:30)
Contemporary tectonic uplift in California and Nevada is an active part of ongoing plate boundary processes driving earthquakes.
However, it has so far been difficult to confidently resolve and interpret uplift patterns. The challenges are twofold. First is the
geodetic problem of isolating the signal of crustal-scale vertical motion given a large number of noisy time series from multiple
networks irregularly distributed in space and time. Second is the problem of partitioning the signals into patterns of long-term
tectonic deformation, earthquake cycle, flexural/isostatic adjustment, local basin response from groundwater hydrology, crustal
loading, and/or mantle flow.
The spread of precise GPS networks and new developments in processing GPS data are leading to a clearer and more complete
picture of vertical motions, improving knowledge of the rates of mountain growth, associated fault slip, and seismic hazard. To
address the explosion in the quantity of new measurements, the ‘Plug and Play’ initiative is easing access to these data for
everyone, enhancing their utility and impact. Plug and Play removes barriers at the beginning and end of the GPS processing
chain, providing a free GPS data processing service and back-end products with free and open access. Currently the system has
global scope, providing products for ~15,000 stations. To derive uplift maps from the data we apply our new robust estimation
method "GPS Imaging" that combines non-parametric robust trend estimation with median-based despeckling and spatial
filtering to suppress noise.
The resulting images show that the Sierra Nevada is the most rapid and extensive uplift feature in the western United States,
rising up to 2 mm/yr, with uplift strongly modulated by climatic and anthropogenic forcing from groundwater pumping. The
images reveal a discontinuity in the uplift field across Owens’ Valley, suggesting that Sierra Nevada uplift is associated with
crustal extension in the southern Walker Lane. Across the Western Transverse Ranges (WTR) of southern California we have taken
the analysis further by combining four geodetic techniques: GPS, InSAR, levelling and tide gauges to constrain the vertical rate
field. These results reveal 1-2 mm/yr of uplift across the WTR and San Gabriel Mountains block, focused west of the San Andreas
fault, consistent with upward interseismic extrusion of these blocks as they experience contraction against the San Andreas Fault.
Offshore Pacific-North America lithospheric structure and Tohoku tsunami observations from a southern
California ocean bottom seismometer experiment, Monica D. Kohler (Caltech)
Tuesday, September 13, 2016 (11:00)
The motivation for the offshore ALBACORE seismic experiment was to identify the physical properties and deformation styles of
the western half of the Pacific-North America plate boundary in southern California. An array of 34 ocean bottom seismometers
(OBSs) was deployed in 2010-2011, extending from the eastern California Borderland into Pacific oceanic plate 300 km west of
the Patton Escarpment. The ALBACORE OBSs, together with 65 stations of the onshore Southern California Seismic Network, were
used to measure ambient noise correlation functions and Rayleigh-wave dispersion curves which were inverted for 3D shear-
wave velocities. The resulting shear-wave velocity model illustrates plate boundary deformation including both thickening and
thinning of the crustal and mantle lithosphere within the California Borderland and at the westernmost edge of the North
American continent. The velocity model defines the transition from continental lithosphere to oceanic tectonic environment, and
indicates the persistence of uppermost mantle volcanic processes associated with East Pacific Rise spreading adjacent to the
Patton Escarpment. One of the most prominent of these seismic structures is a low-velocity anomaly underlying San Juan
Seamount, suggesting ponding of magma at the base of the crust, resulting in thickening and ongoing adjustment of the
WELCOME
2016 SCEC Annual Meeting page 126
lithosphere due to the localized loading. Complementary to this, mapping of two active transpressional fault zones in the
California Borderland - the Santa Cruz-Catalina Ridge fault and the Ferrelo fault - was carried out to characterize their geometries
using over 4500 line-km of new multibeam bathymetry data recorded during the OBS deployment cruise, combined with existing
data. The geometry of the fault systems shows evidence of multiple segments that could experience through-going rupture over
distances exceeding 100 km. Transpression on west- and northwest-trending structures persists as far as 270 km south of the
Transverse Ranges, extension persists in the southern Borderland, and these faults show potential for dip-slip rupture.
The 2011 Tohoku tsunami was serendipitously recorded by the ALBACORE seafloor co-located pressure gauges, demonstrating
how dense array data can illustrate and validate predictions of linear tsunami wave propagation characteristics. Phase and group
travel times were measured for the first arrival in the pressure gauge tsunami waveforms filtered in narrow period bands between
200 and 3000 s. For each period, phase velocities were estimated across the pressure gauge array based on the phase travel-
time gradient using eikonal tomography. Clear correlation is observed between the phase velocity and long-wavelength
bathymetry variations. In the deep open ocean, phase velocity dispersion is observed for both short and long periods. The
pressure gauge tsunami records across the entire array show multiple, large-amplitude, coherent phases arriving one hour to
more than 36 hours after the initial tsunami phase. Beamforming and back-projection analysis of the tsunami waveforms reveals
locations of the bathymetric features in the Pacific Ocean that contributed to the scattered, secondary tsunami arrivals in southern
California.
Plenary Talk Presentations Tuesday
Induced earthquake magnitudes are as large as (statistically) expected, Nicholas J. van der Elst (USGS),
Morgan T. Page (USGS), Debbie A. Weiser (USGS), Thomas H. Goebel (UCSC), & Seyed M. Hosseini (USC)
Tuesday, September 13, 2016 (14:00)
Injection-induced seismicity is a major contributor to seismic hazard in the central US. The question now is whether induced
seismicity is amenable to statistical forecasting on any useful timescale. One of the major uncertainties affecting such a forecast
is how large induced earthquakes can be. Are their maximum magnitudes determined by injection parameters, or by tectonics?
Deterministic limits on induced earthquake magnitudes have been proposed based on the size of the reservoir or the volume of
fluid injected. However, if induced earthquakes occur on tectonic faults oriented favorably with respect to the tectonic stress
field, then they may be limited only by the regional tectonics and connectivity of the fault network. In this study, we show that
the largest magnitudes observed at fluid injection sites are consistent with the sampling statistics of the Gutenberg-Richter
distribution for tectonic earthquakes, assuming no upper magnitude bound. The data pass three specific tests: 1) the largest
observed earthquake at each site scales with the log of the total number of induced earthquakes, 2) the order of occurrence of
the largest event is random within the induced sequence, and 3) the injected volume controls the total number of earthquakes,
rather than the total seismic moment. All three tests point to an injection control on earthquake nucleation, but a tectonic control
on earthquake magnitude. The largest observed earthquakes are exactly as large as expected from the sampling statistics. The
results imply 1) induced earthquake numbers can be estimated from previous activity and anticipated injection volumes, and 2)
induced earthquake magnitudes should be treated with the same maximum magnitude bound that is used to treat seismic
hazard from tectonic earthquakes.
Blurring the boundary between earthquake forecasting and seismic hazard, Craig A. Davis
Tuesday, September 13, 2016 (14:30)
Earthquake forecasting and seismic hazard have been traditionally considered as independent and separate fields of study. We
are currently working on a range of topics that are beginning to blur the bounds between the two fields. With the Canterbury
earthquakes it became apparent that traditional methods of seismic hazard modelling would not fully represent the expected
hazard in the coming 50 years and we developed a hybrid long-term and time-dependent hazard model that melded the two
WELCOME
2016 SCEC Annual Meeting page 127
fields. A distinction of this model is that it combined a range of models that produced forecasts from one-year to 50-years. A
major contribution to the estimated uncertainty in the hazard came from the expected occurrence rate for the region in 20 to
50 years time. We are now developing models to allow improved estimates of this rate and to better capture the epistemic
uncertainty in the short to long-term processes. By using the hybrid model idea, we combine, in an alternative to logic-trees,
models based on differing data sets or hypotheses to produce a forecast of earthquake occurrence. By combining information
from such things as seismicity, geodetic strain, geological data and slow slip events, we can provide forecasts that are typically
more informative than any single model. Additionally, we are trying to better characterize the uncertainty inherent in all of the
models; through an improved understanding of this uncertainty, we can develop better statistical tests of the models, and ideally
provide more useful information to decision makers who are using the outputs of such models both in the form of short-term
earthquake forecasts and long-term hazard forecasts. Finally, through our experiences in Christchurch and more recent
earthquakes in New Zealand, we have gained valuable experience in how to communicate earthquake forecast information and,
also, perhaps, a unique perspective on where future earthquake forecasting and hazard research might be the most beneficial to
end-users of such information.
WELCOME
2016 SCEC Annual Meeting page 128
Plenary Talk Presentation Wednesday
Utilization of earthquake ground motions for nonlinear analysis and design of tall buildings, Gregory
Deierlein (Stanford), Nenad Bijelic (Stanford), & Ting Lin (Marquette) Domniki Asimaki
Wednesday, September 14, 2016 (08:30)
One of the promising applications of simulated ground motions is in the design of tall buildings and other unique structures,
where nonlinear dynamic analyses are used to evaluate their seismic performance. In contrast to current practice, which requires
amplitude scaling of recorded ground motions to represent extreme earthquake hazards, earthquake simulations can enable
more direct assessment without ground motion scaling or reliance on empirical ground motion prediction equations. Moreover,
comprehensive simulations, such as SCEC’s Cybershake study, hold the potential to more realistically capture the influence of
geologic basins and other geologic features on ground motion intensities, frequency content, and durations. This presentation
will briefly review recent research on the use of nonlinear structural analysis in design, with particular focus on characterizing
earthquake ground motions and their effects on structures. The important influence of response spectral shape and ground
motion duration will be illustrated in a case study of a twenty-story building. Efforts to compare and contrast the response of
structures to simulated and record motions will be presented as a step towards evaluating and validating earthquake simulations
for use in engineering practice
Progress Report of the SCEC Utilization of Ground Motion Simulations (UGMS) Committee, C.B. Crouse
(AECOM), & Thomas H. Jordan (USC)
Wednesday, September 14, 2016 (09:00)
The goal of the UGMS committee, since its inception in the spring of 2013, has been to develop digital long-period response
spectral acceleration maps for the Los Angeles region, for inclusion in the NEHRP and ASCE 7 Seismic Provisions and in the Los
Angeles City Building Code. The maps are based on 3-D numerical ground-motion simulations, generated using CyberShake, and
ground motions computed using the four NGA West2 empirical ground-motion prediction equations that incorporate a basin
amplification term. The UGMS developed a method to combine the response spectra from both approaches to generate risk-
targeted Maximum Considered Earthquake (MCER) response spectra for the region. These response spectra cover the period
band from 0 to 10 sec and will be proposed as an amendment to the ASCE -16 standard adopted by the City of Los Angeles.
The MCER response spectrum at a given site would be obtained from a USGS-like web look-up tool. Users would input the site
coordinates and site class (or Vs30), and the output would be a “site-specific” MCER response spectrum and associated SDS and
SD1 values of the Design Response Spectrum. The user would have the option of obtaining more information, such as hazard
curves, risk coefficients, and source and magnitude-distance deaggregation data. The look-up tool is currently being developed
by SCEC with a scheduled release of a trial version toward the end of 2016 or early 2017.
The work of the UGMS committee has been coordinated with (1) the SCEC Ground Motion Simulation Validation Technical
Activity Group (GMSV-TAG), (2) other SCEC projects, such as CyberShake and UCERF, and (3) the USGS national seismic hazard
mapping project.
WELCOME
2016 SCEC Annual Meeting page 129
Poster Session Schedule View full abstracts at www.scec.org/meetings/2016/am
Sunday, September 11, 2016
21:00 – 22:30 Poster Session 1
Monday, September 12, 2016
16:00 – 17:30 Poster Session 2
21:00 – 22:30 Poster Session 3
Tuesday, September 13, 2016
16:00 – 17:30 Poster Session 4
21:00 – 22:30 Poster Session 5
Ground Motion Prediction (GMP) Posters 001-002, 265-280
001 Next Generation SDSU BBP Module Validation, Kim
B Olsen, and Rumi Takedatsu
002 The SCEC Broadband Platform: Open-Source
Software for Strong Ground Motion Simulation and
Validation, Fabio Silva, Christine A Goulet, Philip J
Maechling, Scott Callaghan, and Thomas H Jordan
265 Towards Implementation of Multi-Segment
Ruptures in the Broadband Platform: Composite
Source Model, John G Anderson
266 From VS30 to near-surface velocity profiles:
Integrating soft sediments in SCEC's UCVM,
Domniki Asimaki, Jian Shi, and Ricardo Taborda
267 High-Frequency Nonlinear Simulations of Southern
San Andreas Earthquake Scenarios, Daniel Roten,
Kim B Olsen, Steven M Day, and Yifeng Cui
268 Using Simulated Ground Motions to Constrain
Near-Source Ground Motion Prediction Equations
in Areas Experiencing Induced Seismicity, Samuel A
Bydlon, and Eric M Dunham
269 Analysis of Q-factor’s parameters of Los Angeles
through Simulation and Artificial Intelligence, Naeem
Khoshnevis, and Ricardo Taborda
270 A Flexible and Memory-Efficient Displacement-
Based Approach to Modeling Attenuation in Wave
Propagation, Md Monsurul Huda, Ricardo Taborda, and
Naeem Khoshnevis
271 Ground motions from induced earthquakes in
Oklahoma and Kansas: implications for seismic
hazard, Morgan P Moschetti, Steven Rennolet, Eric M
Thompson, William Yeck, Dan McNamara, Robert
Herrmann, Peter M Powers, and Susan Hoover
272 Verification and Validation of High-Frequency (fmax
= 5 Hz) Ground Motion Simulations of the 2014 M
5.1 La Habra, California, earthquake, Ricardo
Taborda, Kim B Olsen, Robert W Graves, Fabio Silva,
Naeem Khoshnevis, William H Savran, Daniel Roten,
Zheqiang Shi, Christine A Goulet, Jacobo Bielak, Philip J
Maechling, Yifeng Cui, and Thomas H Jordan
273 Ground motion amplification in the Kanto Basin from
future Itoigawa-Shizuoka earthquakes near Tokyo
using virtual earthquakes, Marine A Denolle, Pierre
Boué, Naoshi Hirata, Shigeki Nakagawa, and Gregory C
Beroza
274 Green’s functions retrieved by Multi-Component C3
for Ground Motion Prediction, Yixiao Sheng, Gregory
C Beroza, and Marine A Denolle
275 On nonstationarity corrections and durations in
ground motion applications of random vibration
theory, Chris Van Houtte, Tam Larkin, and Caroline
Holden
WELCOME
2016 SCEC Annual Meeting page 130
276 NGA 2 GMPE’s Under Predict Long-Period Near-
Source Motions from Large Earthquakes, Thomas H
Heaton, and Becky Roh
277 Assessment of Predictive Values of Site Response
based on GMPE approaches using a Large-N array,
Nori Nakata
278 Ground Strains in Southern California from
Earthquakes on the San Jacinto Fault, Andrew J
Barbour, and Annemarie S Baltay
279 Frictional rheology for nonlinear attenuation:
Implications for paleoseismology and strong S-
waves, Norman H Sleep
280 Higher Earthquake Intensity Attenuation Rates in
the Urbanized Southern Puget Lowland Than
Elsewhere Along the Cascadia Forearc, Thomas M
Brocher
Unified Structural Representation (USR) Posters 003-007
003 CFM Version 5.1: New and revised 3D fault
representations and an improved database, Andreas
Plesch, Craig Nicholson, Christopher C Sorlien, John H
Shaw, and Egill Hauksson
005 Structure of the San Andreas Fault Zone in the
Salton Trough Region of Southern California: A
Comparison with San Andreas Fault Structure in the
Loma Prieta Area of Central California, Gary S Fuis,
Rufus D Catchings, Daniel S Scheirer, Mark R Goldman,
and Edward Zhang
006 Anomalous Uplift at Pitas Point: Implications from
Onshore & Offshore 3D Fault & Fold Geometry and
Observed Fault Slip, Craig Nicholson, Christopher C
Sorlien, and Tom E Hopps
007 Displacement direction and 3D geometry for the
south-directed North Channel – Pitas Point fault
system and north-directed ramps, decollements,
and other faults beneath Santa Barbara Channel.,
Christopher C Sorlien, Craig Nicholson, Richard J Behl,
and Marc J Kamerling
Stress and Deformation Over Time (SDOT) Posters 008-024
008 Reconstruction Modeling of Topography and
Lithosphere Dynamics using Western U.S. Strain
History within the Pacific-North America Plate
Boundary Zone, Alireza Bahadori, Bill E Holt, Lucy
Flesch, Lijun Liu, Troy Rasbury, Gavin Piccione, and Rubin
Smith
009 A new way to estimate shear tractions on active
faults in southern California, Peter Bird
010 4D stress evolution models of the San Andreas
Fault System using improved geodetic and
paleoseismic constraints, Bridget R Smith-Konter,
Karen M Luttrell, Xiaopeng Tong, and David T Sandwell
011 Temperature exerts the strongest control on the 3D
rheology of the southern California lithosphere,
Wayne R Thatcher, David S Chapman, and Colin Williams
012 Seasonal water storage modulating seismicity on
California faults, Christopher W Johnson, Yuning Fu,
and Roland Bürgmann
013 Towards Detailed Characterization of Spatio-
temporal Variations in Stress Parameters along the
San Jacinto Fault Zone, Niloufar Abolfathian, Patricia
Martínez-Garzón, and Yehuda Ben-Zion
014 Can tectonic loading be observed as interseismic
stress rotation? Jeanne L Hardebeck
WELCOME
2016 SCEC Annual Meeting page 131
015 Role of fault geometry on the spatial distribution of
the slip budget, Phillip G Resor, Michele L Cooke, Scott
T Marshall, and Elizabeth H Madden
016 Quantifying Late Quaternary deformation in the
Santa Maria Basin: A OSL, GPS and soil
chronosequence based model for determining
strath terrace deformation in the Zaca Creek
drainage, Santa Barbara County, Andrew C Farris,
and Nate W Onderdonk
017 Progress towards deriving the three-dimensional
coseismic deformation field along the White Wolf
fault during the Mw ~7.3 1952 Kern County
earthquake, Alexandra E Hatem, James F Dolan, and
Chris W Milliner
018 Learning viscoelasticity with neural networks,
Phoebe DeVries, Thomas B Thompson, and Brendan J
Meade
019 Using Well Water Level To Measure Volumetric
Strain Considering The Dissipation Effect, Yuqing
Xie, Yonghong Zhao, and Lingsen Meng
020 Coulomb stress evolution over the past 200 years
and seismic hazard along the Xianshuihe fault zone
of Sichuan, China, Zhigang Shao, Jing Xu, Hongsheng
Ma, and Langping Zhang
021 Calculating regional stresses for northern
Canterbury: the effect of the 2010 Darfield
earthquake, Susan Ellis, Charles A Williams, John Ristau,
Martin Reyners, Donna Eberhart-Phillips, and Laura M
Wallace
022 Evidence for strong lateral seismic velocity variation
in the lower crust – upper mantle beneath the
California margin, Voon Hui Lai, Robert W Graves,
Shengji Wei, and Donald V Helmberger
023 Fault-parallel shear fabric in the ductile crust of
Southern California imaged using receiver
functions, Vera Schulte-Pelkum, and Karl Mueller
024 Discriminating Between Induced vs Tectonic
Seismicity in Intraplate Regions: the Contribution of
the Long-Term History of Fault Behavior, Maria
Beatrice Magnani, Michael L Blanpied, Heather DeShon,
and Matthew Hornbach
Fault and Rupture Mechanics (FARM) Posters 025-078
025 Biomarkers as a tool to measure coseismic
temperature rise, Genevieve L Coffey, Heather M
Savage, Pratigya J Polissar, Brett M Carpenter, and
Cristiano Collettini
026 Biomarker thermal maturity at seismic timescales in
high-velocity rotary shear experiments, Hannah S
Rabinowitz, Heather M Savage, Elena Spagnuolo, and
Giulio Di Toro
027 Slip and Seismic Radiation Along Bi-material Faults:
An Experimental Analysis, Brett M Carpenter, Ximeng
Zu, Xiaowei Chen, and Ze'ev Reches
028 Laboratory investigation of friction along rock joints
and identification of peaks in shear stiffness prior to
the joint’s shear failure, Ahmadreza Hedayat, and John
Hinton
029 Earthquake source complexity? Near-fault velocity
spectra from laboratory failures and their relation to
natural ground motion, N. M Beeler, Brian D Kilgore,
and David A Lockner
030 Progress Report on Addition of a High-Speed Drive
to High-Pressure, Rotary-Shear Apparatus, Terry E
Tullis
031 Multi-Scale Flash Heating and Frictional Weakening
at Seismic Slip-Rates in Rock, Frederick M Chester,
Omid Saber, and J. L. Alvarado
WELCOME
2016 SCEC Annual Meeting page 132
032 Dynamic Weakening of Sliding Friction and the
Influence of Gouge Development, Monica R Barbery,
Frederick M Chester, Judith S Chester, and Omid Saber
034 The Scale-Dependence of Fault Roughness and
Asperity Strength, Christopher A Thom, Emily E
Brodsky, and David L Goldsby
035 Laboratory Earthquakes Nucleated by Fluid
Injection, Marcello Gori, Vito Rubino, Ares J Rosakis, and
Nadia Lapusta
036 Estimation of physical properties of a rock sample
based on a laboratory transmitted wave experiment
and 3D numerical simulations, Nana Yoshimitsu, and
Takashi Furumura
037 The Action of water films at Å-Scales in the Earth:
Implications for Midcrustal Eathquakes and
Overpressuring, Kevin M Brown, Dean Poeppe, and
IODP Leg 348 Shipboard Party
038 Physical controls of spontaneous and triggered
slow-slip and stick-slip at the fault gouge scale,
Behrooz Ferdowsi, and David L Goldsby
039 Dynamic friction in sheared fault gouge: implications
of acoustic vibration on triggering and slow slip, Jean
M Carlson, Charles K Lieou, and Ahmed E Elbanna
040 Localization and instability in sheared granular
materials: Role of friction and vibration, Konik R
Kothari, and Ahmed E Elbanna
041 How we learned to stop worrying and start loving
bulk nonlinearities, Setare Hajarolasvadi, and Ahmed E
Elbanna
042 Brittle to ductile transition in a model of sheared
granular materials, Xiao Ma, and Ahmed E Elbanna
043 Flexible implementation of multiphysics and
discretizations in PyLith crustal deformation
modeling software, Brad T Aagaard, Charles A Williams,
and Matthew G Knepley
044 Earthquake cycles on rate-state faults: how does
recurrence interval and its variability depend on fault
length? Camilla Cattania, and Paul Segall
045 Plasticity Throughout the Earthquake Cycle, Brittany
A Erickson, Eric M Dunham, and Jeremy E Kozdon
046 Localization and delocalization of shear in fault
gouge from thermal pressurization, Shanna Chu, and
Eric M Dunham
047 Aiming for Validation – The SCEC/USGS Dynamic
Earthquake Rupture Code Comparison Exercise,
Ruth A Harris
048 Spontaneous Dynamic Rupture Simulation on
Geometrically Complex Faults Governed by
Different Friction Laws, Bin Luo, and Benchun Duan
049 Linking grain-scale and fault-scale earthquake
simulations, Ryan Payne, Benchun Duan, and David
Sparks
050 Stress heterogeneity at restraining double bends
under multicycles and its effect on rupture
propagation in 3D, Dunyu Liu, and Benchun Duan
051 Modeling the effect of roughness on the nucleation
and propagation of shear rupture, Yuval Tal, and
Bradford H Hager
052 Friction and stress drop controlled by fault
roughness and strength heterogeneity, Olaf Zielke,
Martin Galis, and Paul M Mai
053 Comparison of Source Inversions and Stress Drops
with In-Situ Observations of Faulting, Pamela A
Moyer, Margaret S Boettcher, and William L Ellsworth
054 Natural polish in granitic rocks, Shalev Siman-Tov,
Emily E Brodsky, and Greg M Stock
055 Asperity break after 12 years: The Mw6.4 2015
Lefkada (Greece) earthquake, Frantisek Gallovic,
Efthimios Sokos, Jiri Zahradnik, Anna Serpetsidaki,
Vladislav Plicka, and Anastasia Kiratzi
WELCOME
2016 SCEC Annual Meeting page 133
056 A possible joint San Andreas-Imperial fault rupture
in the Salton Trough region, David D Oglesby,
Christodoulos Kyriakopoulos, Aron J Meltzner, and
Thomas K Rockwell
057 Investigating the high-frequency, weak-motion
nucleation phase initiating the June 10th, 2016 Mw
5.2 Borrego Springs Earthquake, Adam Arce, Mareike
N Adams, and Chen Ji
058 Estimates of fault tractions in the San Gorgonio
Pass region consider fault interaction, Aviel R Stern,
Michele L Cooke, Jennifer L Beyer, Jennifer M Tarnowski,
David D Oglesby, and Roby Douilly
059 Development of extensional step overs within
anisotropic systems: Implications for the Rodgers
Creek-Hayward step over, Jessica A McBeck, and
Michele L Cooke
060 Sensitivity of deformation to activity along the Mill
Creek strand of the San Andreas fault within the San
Gorgonio Pass, Jennifer L Beyer, and Michele L Cooke
061 Dynamic Models of Large Ruptures on the Southern
San Andreas Fault, Julian C Lozos
062 Off-fault plasticity in dynamic rupture simulations:
3D numerical analysis and effects on rupture
transfer in complex fault geometries, Stephanie
Wollherr, and Alice-Agnes Gabriel
063 Earthquake source parameter validation using
multiple-scale approaches for induced seismicity in
Oklahoma, Rachel E Abercrombie, and Xiaowei Chen
064 Scaling of finite-source parameters at Parkfield,
California, Katie E Wooddell, Doug S Dreger, Taka'aki
Taira, Robert M Nadeau, and Luca Matagnini
065 Geodetic Measurements of Slow Slip and Tremor in
Parkfield, CA, Brent G Delbridge, Roland Bürgmann,
and Robert M Nadeau
066 Reconciling seismicity and geodetic locking depths
on the Anza segment of the San Jacinto Fault, Junle
Jiang, and Yuri Fialko
067 Rupture evolution and high-frequency radiation
during mega/large earthquakes, resolved by hybrid
backprojection technique, Ryo Okuwaki, and Yuji Yagi
068 Self-similar asymptotics of rate-strengthening faults,
Robert C Viesca, and Pierre Dublanchet
069 Coseismic Strengthening of the Shallow Subduction
Megathrust Further Enhances Inelastic Wedge
Failure and Efficiency of Tsunami Generation, Evan
T Hirakawa, and Shuo Ma
070 How does fault geometry control earthquake
magnitude? Quentin Bletery, Amanda Thomas, Leif
Karlstrom, Alan W Rempel, Anthony Sladen, and Louis
De Barros
071 Dehydration-Induced Porosity Waves and Episodic
Tremor and Slip, Robert Skarbek, and Alan W Rempel
072 Tremor and Slow Slip in the West Antarctic Ice
Sheet, Bradley P Lipovsky, and Eric M Dunham
073 Modern earthquake ruptures on contrasting fault
systems in the Laguna Salada region of
northwestern Mexico, Ronald M Spelz, John M
Fletcher, Thomas K Rockwell, Alejandro Hinojosa,
Michael E Oskin, Lewis A Owen, Jaziel F Cambron, Víctor
H Villaverde, Laura V Vallin, Abel A Gutierrez, and Keene
W Karlsson
074 The mechanics of multifault ruptures and the
keystone fault hypothesis, John M Fletcher, Michael E
Oskin, and Orlando J Teran
075 Effects of elastoplastic material properties on
shallow fault slip and surface displacement fields,
Johanna M Nevitt, Benjamin A Brooks, Todd L Ericksen,
Craig L Glennie, Christopher M Madugo, David A
Lockner, Diane E Moore, Sarah E Minson, and Kenneth
W Hudnut
WELCOME
2016 SCEC Annual Meeting page 134
076 Kinematic Rupture Process of the 2016 Mw 7.1
Kumamoto Earthquake Sequence, Han Yue, Mark
Simons, Cunren Liang, Heresh Fattahi, Eric J Fielding,
Hiroo Kanamori, Donald V Helmberger, Linjun Zhu,
Michael Sylvain, and Jean-Philippe Avouac
077 Shallow level incipient pulverization found with a
restraining bend of the Clark segment of the San
Jacinto Fault, southern California, Daniel W Peppard,
Heather N Webb, and Gary H Girty
078 Post-seismic ductile flow beneath the brittle-plastic
transition: constraints on rheology from
microstructural and electron backscatter diffraction
(EBSD) analyses of mylonitized pseudotachylite,
South Mountains metamorphic core complex,
Arizona, Craig Stewart, and Elena Miranda
WELCOME
2016 SCEC Annual Meeting page 135
Earthquake Geology Posters 079-115
079 Fault Scarp Degradation Analysis At Dragon’s Back
Using High Resolution Topography Data, Emil
Chang, Gilles Peltzer, and Seulgi Moon
080 Cumulative offsets revealed by airborne LiDAR
along a “creeping” section of the Haiyuan fault,
northern Tibetan plateau, Jing Liu, Tao Chen, Yanxiu
Shao, Peizhen Zhang, Kenneth W Hudnut, Qiyun Lei, and
Zhanfei Li
081 New Investigations on the Hollywood-Raymond
Fault Zone, Los Angeles, California, Janis Hernandez,
Rufus D Catchings, Robert R Sickler, and Mark R
Goldman
082 Structural Architecture of the Western Transverse
Ranges and Potential for Large Earthquakes, Yuval
Levy, and Thomas K Rockwell
083 Marine terraces, isostasy, earthquakes, and uplift
rates, Alexander R Simms, Helene Rouby, Kurt Lambeck,
and Angela Roman
084 Geology of liquefaction-induced lateral spreading,
new results from Christchurch, New Zealand,
Gregory P De Pascale, Jeffrey L Bachhuber, Ellen Rathje,
Jing Hu, Peter Almond, Christian Ruegg, and Mike
Finnemore
085 Characterizing emissivity spectra from geomorphic
surfaces along the southern San Andreas Fault,
Ryan D Witkosky, Paul M Adams, Kerry Buckland,
Kenneth W Hudnut, Patrick D Johnson, David K Lynch,
Katherine M Scharer, Joann M Stock, and David M Tratt
086 Paleoseismic investigation of the Rose Canyon fault
zone, San Diego, CA, Eui-jo D Marquez, Jillian M
Maloney, Thomas K Rockwell, Neal W Driscoll, Scott
Rugg, and Jeff M Babcock
087 Evidence for Abrupt Subsidence Event in
Carpinteria Marsh at 2 ± 0.2 ka, Laura C Reynolds,
Alexander R Simms, Thomas K Rockwell, John M Bentz,
and Robert B Peters
088 Aerial2lidar3d: A New Point Cloud-Optical Image
Matching Technique to Quantify Near-Field,
Surface Co-Seismic Deformation in 3D: Application
to the 1999 Mw 7.1 Hector Mine Earthquake and
New Surface Offset Measurements of the 1999 Mw
7.6 Izmit earthquake, Chris W Milliner, James F Dolan,
Robert Zinke, James Hollingsworth, Sebastien Leprince,
and Francois Ayoub
089 Preliminary results on Late Quaternary slip rate of
the central Haiyuan Fault constrained by terrestrial
in situ cosmogenic nuclides dating, UAV and LiDAR
surveys, Yanxiu Shao, Jing Liu, Michael E Oskin, Jerome
Van der Woerd, Jinyu Zhang, Peng Wang, Pengtao
Wang, Wang Wang, and Wenqian Yao
090 Published Quaternary geochronologic data show
the importance of dating geomorphic surfaces with
multiple geochronometers, Peter O Gold, and Whitney
M Behr
091 Site-specific earthquake hazard characterization for
Dhaka City, Bangladesh, Mir F Karim, Zillur M
Rahman, Maksud Kamal, and Sumi Siddiqua
092 Late Holocene rupture history of the Ash Hill fault,
Eastern California Shear Zone, Christine Regalla,
Hannah Pangrcic, Eric Kirby, and Eric McDonald
093 San Diego Earthquake Hazard: Geotechnical Data
Synthesis, Luke Weidman, Jillian M Maloney, and
Thomas K Rockwell
094 Quaternary Expression of Northern Great Valley
Faults and Folds: Accommodating North-South
Contraction in the Northeastern California Shear
Zone, Stephen J Angster, Thomas Sawyer, and Steven G
Wesnousky
095 Geological Observations on History and Future of
Large Earthquakes along the Himalayan Frontal
Fault Relative to the April 25, 2015 M7.8 Gorkha
Earthquake near Kathmandu, Nepal, Steven G
Wesnousky
WELCOME
2016 SCEC Annual Meeting page 136
096 Late Quaternary slip rates from offset alluvial fan
surfaces along the Central Sierra Madre fault,
southern California, Austin Hanson, Reed J Burgette,
Katherine M Scharer, and Nikolas Midttun
097 New slip rates and characterization of active faults
in the northern Walker Lane, Ian Pierce, Steven G
Wesnousky, and Lewis A Owen
098 Reevaluation of high slip rates on the Eglington
Fault Las Vegas, NV utilizing new
chronostratigraphic and geologic evidence, Kathleen
B Springer, and Jeffrey S Pigati
099 Testing for slip rate changes on the Sierra Madre
fault: Progress on dating an offset terrace surface of
possible middle Pleistocene age, Reed J Burgette,
Nathaniel Lifton, Katherine M Scharer, Devin McPhillips,
and Austin Hanson
100 Lidar mapping and luminescence dating reveal
highly variable latest Pleistocene-Holocene slip
rates on the Awatere fault at Saxton River, South
Island, New Zealand, Robert Zinke, James F Dolan,
Russ J Van Dissen, Ed J Rhodes, and Christopher P
McGuire
101 Progress towards a comprehensive incremental slip
rate and paleo-earthquake age and displacement
record for the central Garlock fault, James F Dolan,
Sally F McGill, Ed J Rhodes, and Thomas M Crane
102 Observations from preliminary 1:24,000 scale
bedrock mapping in the western Joshua Tree
region: Potential connections between historic
seismicity and evolving fault systems, Ann Hislop,
Robert E Powell, Sean P Bemis, David P Moecher, and
Luke C Sabala
104 Towards morphologic and cosmogenic dating of
paleoearthquake and fault slip rates in the Central
Nevada Seismic Belt, Tabor J Reedy, and Steven G
Wesnousky
105 Investigate fault zone hydrogeologic architectures
by using water level tidal and barometric response,
Lian Xue, Emily E Brodsky, Vincent Allegre, Patrick M
Fulton, L. Parker L Beth, and John A Cherry
106 Floods, storms, and the identification of wave-
dominated deltas: Insights from ground-penetrating
radar profiles of the Oxnard Plain, Julie M Zurbuchen,
and Alexander R Simms
107 Fault Deformation and Segmentation of the
Newport-Inglewood Rose Canyon, and San Onofre
Trend Fault Systems from New High-Resolution 3D
Seismic Imagery, James J Holmes, Neal W Driscoll, and
Graham M Kent
108 Geometric, kinematic, and temporal patterns of
Quaternary surface rupture on the Eastern Pinto
Mountain fault zone near Twentynine Palms,
southern California, Christopher M Menges, Jonathan
C Matti, and Stephanie L Dudash
109 Assessing evidence for connectivity between the
San Diego Trough and San Pedro Basin fault
systems, offshore Southern California, Jayne M
Bormann, Graham M Kent, Neal W Driscoll, and Alistair
J Harding
110 Multi-scale Structural Characterization of the Mecca
Hills Fault System in the NE block of the Southern
San Andreas Fault System, California, Kelly K
Bradbury, Amy C Moser, Sarah A Schulthies, and James
P Evans
111 Imaging fault scarps and fault zone evolution near
an oceanic transform fault using high-resolution
bathymetry, Curtis W Baden, George E Hilley, Samuel
Johnstone, Robert M Sare, Felipe Aron, Holly Young,
Christopher M Castillo, Lauren Shumaker, Johanna M
Nevitt, Tim McHargue, and Charles Paull
113 Seafloor expression of active transpressional
faulting offshore Southern California, Mark R Legg,
Simon L Klemperer, Christopher M Castillo, Marie-
Helene Cormier, Michael Brennan, Katy Croff Bell, Dwight
Coleman, Chris Goldfinger, and Jason Chaytor
WELCOME
2016 SCEC Annual Meeting page 137
114 Surface slip behavior of the southern and central
Alpine fault, New Zealand, Jozi K Pearson, and Nicolas
C Barth
115 Rupture zone characteristics of the 2010 El Mayor-
Cucapah (Mexico) earthquake revealed with
differential lidar, Lia J Lajoie, and Ed Nissen
Southern San Andreas Fault Evaluation (SoSAFE) Posters 116-131
116 The East Shoreline strand of the San Andreas Fault
and its implications for the next Big One in southern
California, Susanne U Janecke, Daniel Markowski, Roger
Bilham, James P Evans, Michael Bunds, Jack Wells,
Jeremy Andreini, and Robert Quinn
117 Applying newly developed luminescence dating to
alluvial fans in the Anza Borrego Desert, southern
California, Brittney Emmons, Seulgi Moon, Nathan D
Brown, Kimberly D Blisniuk, and Ed J Rhodes
118 New Holocene slip-rate sites along the Mojave San
Andreas Fault near Palmdale, CA, Elaine K Young,
Eric S Cowgill, and Katherine M Scharer
119 Preliminary late Pleistocene slip rate for the western
Pinto Mountain fault, Morongo Valley, southern
California, Katherine Gabriel, Doug Yule, and Richard V
Heermance
120 High-resolution imaging of the San Andreas Fault
around San Gorgonio Pass using fault zone head
waves and double-difference tomography, with
implications for large earthquake ruptures, Pieter-
Ewald Share, Yehuda Ben-Zion, Clifford H Thurber,
Haijiang Zhang, and Hao Guo
121 Comparison of the rupture history of the southern
San Andreas fault with empirical data on fault
displacement and rupture length, Katherine M Scharer
122 Sedimentary provenance constraints on the
Quaternary faulting history of the Mission Creek
fault strand, southern San Andreas Fault Zone, CA,
Julie C Fosdick, Kimberly D Blisniuk, and Louis Wersan
123 Geologic framework of the El Casco 7.5’
quadrangle: southwestern gateway to the San
Gorgonio Pass knot in the San Andreas Fault zone,
Jonathan C Matti
124 Slip variability and temporal clustering along the
Imperial fault at Mesquite Basin, Imperial Valley,
California, and possible through-going rupture to the
San Andreas fault, Aron J Meltzner, Thomas K
Rockwell, Rebecca Y Tsang, and Paula M Figueiredo
125 Work in progress to estimate a Latest Pleistocene
slip rate for the Banning Strand of the San Andreas
Fault near North Palm Springs, Sally F McGill, Paula
M Figueiredo, and Lewis A Owen
126 Isochron burial dating of paleosols within the
Whitewater Fan, northern Coachella Valley,
California, Nathaniel Lifton, Richard V Heermance,
Doug Yule, and Brittany Huerta
127 Holocene geologic slip rate for the Mission Creek
fault at the Three Palms site in the Indio Hills, Juan J
Munoz, Whitney M Behr, Warren D Sharp, Rosemarie
Fryer, and Peter O Gold
128 Testing the shorter and variable recurrence interval
hypothesis along the Cholame segment of the San
Andreas Fault, Alana M Williams, Ramon Arrowsmith,
Sinan O Akciz, Thomas K Rockwell, Lisa Grant Ludwig,
and Sally Branscomb
129 Mapping fault creep, frictional properties, and
unrecognized active structures with dense geodetic
data in the Imperial Valley, Southern California, Eric
O Lindsey, and Yuri Fialko
130 Investigating the age and origin of small offsets at
Van Matre Ranch along the San Andreas Fault in
the Carrizo Plain, California, James B Salisbury,
Ramon Arrowsmith, Thomas K Rockwell, Sinan O Akciz,
Nathan D Brown, and Lisa Grant Ludwig
131 Activity of the Mill Creek and Mission Creek strands
of the San Andreas fault through the San Gorgonio
WELCOME
2016 SCEC Annual Meeting page 138
Pass region, Alex E Morelan, Michael E Oskin, Judith S
Chester, and Daniel F Elizondo
Tectonic Geodesy Posters 132-168
132 Interpreting Oligocene Paleogeography in Southern
California Using a Provenance Analysis of the
Sespe Formation, Carl Swindle, and Parker Clarke
133 High-Resolution Topographic Mapping of Active
Faults in Southern California with Satellite Optical
Imagery, William D Barnhart, Michael J Willis, Terryl L
Bandy, Rich Briggs, Brianna Morales, and Mark Faney
134 Surface slip rate of the Imperial Fault estimated from
remote controlled quadcopter photogrammetry,
John DeSanto, and David T Sandwell
135 Seventy-two years of Surface Creep on the North
Anatolian fault at Ismetpasa: Implications for the
southern San Andreas and Hayward faults, Roger
Bilham, Haluk Ozener, David J Mencin, Asli Dogru, Semih
Ergintav, Ziyadin Cakir, Alkut Aytun, Bahadir Aktug, Onur
Yilmaz, Wade Johnson, and Glen S Mattioli
136 Fault creep observed on the Maacama and Rodgers
Creek faults, northern California using PS-InSAR,
Jerlyn L Swiatlowski, and Gareth J Funning
137 InSAR observations in Southern California, Rowena
B Lohman, and Kyle D Murray
138 Characterization of fault motions observed with
UAVSAR, Jay W Parker, Andrea Donnellan, Scott
Hensley, and Margaret T Glasscoe
139 Toward the 3-component Crustal Motion Model:
Integration of Sentinel-1A SAR interferometry and
continuous GPS in the Los Angeles-Western
Mojave area, Ekaterina Tymofyeyeva, Homan Lau, and
Yuri Fialko
140 Combination of GPS and InSAR Data for Crustal
Deformation Mapping, Zhen Liu, Zheng-Kang Shen,
and Cunren Liang
141 The SCEC Community Geodetic Model V1:
Horizontal Velocity Grid, David T Sandwell, Yuehua
Zeng, Zheng-Kang Shen, Brendan W Crowell, Jessica R
Murray, Robert McCaffrey, and Xiaohua Xu
142 Steady and time-dependent strain rate maps of
California from inversion of GPS time series, Robert
McCaffrey
143 Comparison of GPS Strain Rate Computing
Methods and Their Application, Yanqiang Wu
144 The California Plate Boundary Observatory GPS-
GNSS Network, Christian Walls, Doerte Mann, Ryan C
Turner, Andre Basset, Shawn Lawrence, Kenneth Austin,
Stephen T Dittman, Karl Feaux, and Glen S Mattioli
145 USGS Southern California GPS Network, Daniel N
Determan, Aris G Aspiotes, Derik T Barseghian, Kenneth
W Hudnut, and Keith F Stark
146 Guadalupe Island as a constrain for earthquake
hazard at Californias's shoreline, Jose Javier
Gonzàlez-Garcìa, Alejandro Gonzalez-Ortega, John E
Galetzka, Christian Walls, Hebert Martinez-Barcena, and
Juan C Robles
147 Geodetic slip rate estimates in California, and their
uncertainties, Eileen L Evans
148 Strain accumulation on faults beneath Los Angeles:
a geodesy-based picture accounting for the effects
of sedimentary basins and anthropogenic surface
deformation, Chris Rollins, Donald F Argus, Walter
Landry, Sylvain D Barbot, and Jean-Philippe Avouac
149 New GPS Site Velocities in San Gorgonio Pass,
Southern California and Preliminary Elastic
Modeling, Naomi Jahan, Matthew Peterson, Sally F
McGill, and Joshua C Spinler
WELCOME
2016 SCEC Annual Meeting page 139
150 Vital Signs of the Planet: Southern California
Education Contribute to Crustal Deformation
Studies Within San Bernardino and Riverside
Counties, Dan Keck
151 Is the CFM5.0 an Improvement? Evidence from
Mechanical Models of the Western Transverse
Ranges Region, Scott T Marshall, Gareth J Funning,
and Susan E Owen
152 Constraining Interacting Fault Models in the Salton
Trough with Remote Sensing Data, Margaret T
Glasscoe, Jay W Parker, Andrea Donnellan, Gregory A
Lyzenga, and Chris W Milliner
153 Single station automated detection of transient
deformation in GPS time series with the relative
strength index: A case study of Cascadian slow-slip,
Brendan W Crowell, Yehuda Bock, and Zhen Liu
154 Estimating secular velocities from GPS data
contaminated by postseismic motion at sites with
limited pre-earthquake data, Jessica R Murray, and
Jerry Svarc
155 Geodetic moment accumulation and seismic
release in northern Baja California, Mexico,
Alejandro Gonzalez-Ortega, Jose Javier Gonzàlez-Garcìa,
and David T Sandwell
156 Postseismic deformation and viscosity of the
Mexicali region following the El-Mayor Cucapah
earthquake inferred from GPS observations,
Xiaopeng Tong, Alejandro Gonzalez-Ortega, Bridget R
Smith-Konter, and David T Sandwell
157 Coseismic and postseismic deformation from the
2010 Mw 7.2 El Mayor-Cucapah Earthquake in Baja
California: Lithospheric structure and deformation in
the Salton Trough, Mong-Han Huang, Haylee
Dickinson, Andrew Freed, Eric J Fielding, Roland
Bürgmann, and Alejandro A Gonzalez-Ortega
158 Observations and Models of Co- and Post-Seismic
Deformation Due to the 2015 Mw 7.8 Gorkha
(Nepal) Earthquake, Kang Wang, and Yuri Fialko
159 Toward Rapid, Robust Characterization of
Subduction Zone Earthquakes: Application to the
2015 Illapel, Chile Earthquake, Tisha Irwin, and
Gareth J Funning
160 The aseismic slip acceleration before the 2015
Illapel Mw 8.4 event from repeating earthquake
observations, Hui Huang, and Lingsen Meng
161 Using deformation rates in Northern Cascadia to
constrain time-dependent stress- and slip-rate on
the megathrust, Lucile Bruhat, and Paul Segall
162 Compaction-induced elevated pore pressure and
creep pulsing in California faults, Mostafa
Khoshmanesh, and Manoochehr Shirzaei
163 Surface uplift and time-dependent seismic hazard
due to fluid-injection: Cast studies from texas,
Manoochehr Shirzaei
164 InSAR MSBAS Time-Series Analysis of Induced
Seismicity in Cushing, Oklahoma, Magali Barba, and
Kristy F Tiampo
165 Tracking Seasonal Influences on Stress Changes
using Estimates of Anomalous Geodetic Strains, Bill
E Holt, Meredith L Kraner, Adrian A Borsa, and Alireza
Bahadori
166 Using GPS Imaging to Resolve Seasonal Strain in
Central California, Meredith L Kraner, William C
Hammond, Corné W Kreemer, and Geoffrey Blewitt
167 Seasonal motions and interseismic strain measured
by continuous GPS throughout the Transverse
Ranges, CA, Hannah E Krueger, Scott T Marshall, Susan
E Owen, Gareth J Funning, and Jack P Loveless
168 Tectonic and Anthropogenic Deformation
Surrounding the Cerro Prieto Geothermal Field from
Sentinel-1 Interferometry, Xiaohua Xu
WELCOME
2016 SCEC Annual Meeting page 140
Seismology Posters 169-263
169 Aftershock productivity of large megathrust
earthquakes: regional variations and influence of
mainshock source parameters, Nadav Wetzler, Emily
E Brodsky, and Thorne Lay
170 Shear Wave Velocities in the Central Eastern United
States, Mehrdad Hosseini, Paul G Somerville, Andreas
Skarlatoudis, Jeff R Bayless, and Hong Kie Thio
171 Were the 1952 Kern County and 1933 Long Beach,
California, Earthquakes Induced? Susan E Hough,
Victor C Tsai, Robert L Walker, Morgan T Page, and
Seyed M Hosseini
172 Seismic Risk from Induced and Natural
Earthquakes for the Central and Eastern United
States, Taojun Liu, Nicolas Luco, and Abbie B Liel
173 Local near-instantaneously dynamically triggered
aftershocks of large earthquakes, Wenyuan Fan, and
Peter M Shearer
174 A detailed automatic 1998-2015 earthquake catalog
of the San Jacinto fault zone region, Malcolm C
White, Zachary E Ross, Frank L Vernon, and Yehuda Ben-
Zion
175 Reducing Uncertainty in Ground Motion Prediction
Equations by Understanding Path Effects, Valerie J
Sahakian, Annemarie S Baltay, and Tom C Hanks
176 Fault Damage Zone of the 2014 Mw 6.0 South Napa
Earthquake, California, Viewed from Fault-Zone
Trapped Waves, Yong-Gang Li, Rufus D Catchings, and
Mark R Goldman
177 Topographic Influence on Near-Surface Seismic
Velocity in southern California, Jessica C Lin, Seulgi
Moon, Lingsen Meng, and Paul M Davis
178 Validation of Deterministic Broadband Ground
Motion and Variability with Ground Motion, Kyle B
Withers
179 Investigating tremor sources along the San Andreas
Fault using integrated static and dynamic stress
models, Hector Gonzalez-Huizar, Sandra Hardy, and
Bridget Smith-Konter
180 Science-based decision making in a high-risk
energy production environment, Debbie Weiser
181 Effects of regulation on induced seismicity in
southern Kansas, Justin L Rubinstein, William L
Ellsworth, and Sara L Dougherty
182 Data-driven ambient noise correlation for
characterization and fragility evaluation in power
grids, Qianli Chen, and Ahmed E Elbanna
183 The Origin of High-angle Dip-slip Earthquakes at
Geothermal Fields in California, Martin Schoenball,
Patricia Martínez-Garzón, Andrew J Barbour, and
Grzegorz Kwiatek
184 Seismicity and spectral analysis in Salton Sea
Geothermal Field, Yifang Cheng, and Xiaowei Chen
185 Internal structure of the San Jacinto fault zone at
Blackburn Canyon from a dense linear deployment
across the fault, Yehuda Ben-Zion, Pieter-Ewald Share,
Amir A Allam, Fan-Chi Lin, and Frank L Vernon
186 Body wave phases from higher order cross-
correlation, Zack Spica, and Gregory C Beroza
187 ShakeAlert Testing and Certification Platform: Point
Source and Ground Motion Based Evaluations,
Elizabeth S Cochran, Monica D Kohler, Douglas D Given,
Jennifer Andrews, Men-Andrin Meier, Egill Hauksson,
Sarah E Minson, Mohammad Ahmad, Jonathan DeLeon,
and Stephan Guiwits
188 Estimating amplitude uncertainty for normalized
ambient seismic noise cross-correlation with
examples from southern California, Xin Liu, Gregory
C Beroza, and Yehuda Ben-Zion
189 Proximity of Precambrian basement affects the
likelihood of induced seismicity in the Appalachian,
WELCOME
2016 SCEC Annual Meeting page 141
Illinois, and Williston basins, Rob Skoumal, Michael
Brudzinski, and Brian Currie
190 Monitoring of Microseismicity in the Peach Tree
Valley Region with Array Techniques, Jose Luis L
Garcia-Reyes, and Robert W Clayton
191 Two decades of shear-wave splitting
measurements in southern California, Zefeng Li, and
Zhigang Peng
192 Smartphone-Based Earthquake Early Warning in
Chile, Sarah E Minson, Benjamin A Brooks, Sergio
Barrientos, Juan Carlos Baez, Maren Boese, Todd L
Ericksen, Christian Guillemot, Elizabeth S Cochran, Jessica
R Murray, John O Langbein, Craig L Glennie, and
Christopher Duncan
193 Precise relative stress drops, Bruce E Shaw, William L
Ellsworth, Nana Yoshimitsu, Yihe Huang, and Gregory C
Beroza
194 Earthquakes Induced by Hydraulic Fracturing and
Wastewater Injection in Guy-Greenbrier, Arkansas,
Clara E Yoon, Yihe Huang, William L Ellsworth, and
Gregory C Beroza
195 Dynamic stress triggering in the 2010-2011
Canterbury earthquake sequence, Caroline Holden,
Charles A Williams, and David A Rhoades
196 A Simulation Approach for Better Understanding of
Seismic Hazard and Risk in Montreal, Yelena
Kropivnitskaya, Kristy F Tiampo, Jinhui Qin, and Michael
Bauer
197 LArge-n Seismic Survey in Oklahoma (LASSO):
Probing injection-induced seismicity with a dense
array, Sara L Dougherty, Elizabeth S Cochran, and
Rebecca M Harrington
198 Products and Services Available from the Southern
California Earthquake Data Center (SCEDC) and
the Southern California Seismic Network (SCSN),
Ellen Yu, Prabha Acharya, Aparna Bhaskaran, Shang-Lin
Chen, Jennifer Andrews, Valerie Thomas, Egill Hauksson,
and Robert W Clayton
199 Rayleigh-wave ellipticity obtained from noise cross-
correlations across southern California, Elizabeth
Berg, Fan-Chi Lin, and Amir A Allam
200 GrowClust: A hierarchical clustering algorithm for
relative earthquake relocation, with application to
the Spanish Springs and Sheldon, Nevada,
earthquake sequences, Daniel T Trugman, Peter M
Shearer, and Ken D Smith
201 Constraints on Fault Damage Zone Properties and
Normal Modes from a Dense Linear Array
Deployment along the San Jacinto Fault Zone, Amir
A Allam, Fan-Chi Lin, Pieter-Ewald Share, Yehuda Ben-
Zion, Frank L Vernon, Gerard Schuster, and Marianne
Karplus
202 The 2016 Mw5.1 Fareview, Oklahoma earthquakes:
Evidence for long-range poroelastic triggering at
>30 km from disposal wells, Thomas H Goebel,
Matthew Weingarten, Xiaowei Chen, Jackson Haffener,
and Emily E Brodsky
203 Investigating Complex Slow Slip Evolution with
High-Resolution Tremor Catalogs and Numerical
Simulations, Yajun Peng, and Allan M Rubin
204 Constraints on residual topography and crustal
properties in the western US from virtual deep
seismic sounding, Chunquan Yu, Wang-Ping Chen, and
Rob van der Hilst
205 Using coda waves to resolve the scattering and
intrinsic attenuation structure of Southern California,
Wei Wang, and Peter M Shearer
206 Evaluation of Site Response in the Los Angeles
Basin from Spectral Ratio Analysis of Microtremor
Data from a High Density Temporary Broadband
Deployment, Raymond Ng, and Jascha Polet
207 Nowcasting Oklahoma and The Geysers, California,
Molly Luginbuhl, John B Rundle, and Donald L Turcotte
WELCOME
2016 SCEC Annual Meeting page 142
208 Mitigating the spatial biases of back-projection
imaging using a “station-based” slowness
calibration, Han Bao, Lingsen Meng, and Zhipeng Liu
209 Ground Motion Amplitude Variations in the Los
Angeles Region Measured by the Community
Seismic Network, Robert W Clayton, Robert Sanchez,
Julian Bunn, and Richard Guy
210 Observations of the Temporal Evolution of
Earthquake Ruptures, Men-Andrin Meier
211 Rapid Rupture Directivity Determination of
Moderate Dip-Slip Earthquakes with the Reduced
Finite Source Approximation, Xiaohui He, and Sidao
Ni
212 Evidence for Non-Self-Similarity of
Microearthquakes Recorded at a Taiwan Borehole
Seismometer Array, Yen-Yu Lin, Kuo-Fong Ma, Hiroo
Kanamori, Teh-Ru A Song, Nadia Lapusta, and Victor C
Tsai
213 Development and validation of an improved seismic
velocity model for the San Jacinto Fault region of
Southern California, Clifford H Thurber, Amir A Allam,
Xiangfang Zeng, Yan Luo, Hongjian Fang, and Haijiang
Zhang
214 Improved Understanding of Triggered Seismicity in
the Salton Sea Geothermal Field with Waveform
Matching Method, Chenyu Li, Zhigang Peng,
Dongdong Yao, Bridget Casey, and Xiaofeng Meng
215 Tremor and LFE activities using mini seismic array
in the Alaska-Aleutian subduction zone, Bo LI, and
Abhijit Ghosh
216 Comprehensive Detection and Relocation of the
2010 Mw7.2 El Mayor-Cucapah Foreshock
Sequence, Dongdong Yao, Zhigang Peng, Xiaofeng
Meng, Xiaowei Chen, Raul R Castro, and Sizuang Deng
217 A new strategy for earthquake focal mechanisms
using waveform-correlation-derived relative
polarities and cluster analysis: Application to a fluid-
driven earthquake swarm, David R Shelly, Jeanne L
Hardebeck, William L Ellsworth, and David P Hill
218 Complex active faults and seismicity rates in the
trifurcation area of the San Jacinto fault zone
illuminated by aftershocks of the 2016 Mw 5.2
Borrego Springs earthquake, Zachary E Ross, Egill
Hauksson, and Yehuda Ben-Zion
219 Internal structure of the San Jacinto fault zone at
Jackass Flat from data recorded by a dense linear
array, Hongrui Qiu, Yehuda Ben-Zion, Zachary E Ross,
Pieter-Ewald Share, and Frank L Vernon
220 Internal structure of the San Jacinto fault zone in the
trifurcation area southeast of Anza, California, from
data of dense linear arrays, Lei Qin, Yehuda Ben-Zion,
Hongrui Qiu, Pieter-Ewald Share, Zachary E Ross, and
Frank L Vernon
222 Testing and Reconciling Stress Drop and
Attenuation Models in Southern California, Peter M
Shearer, Rachel E Abercrombie, and Daniel T Trugman
223 Characterizing the Geometry and Seismotectonics
of the Hilton Creek Fault System, Kyle P Macy, Amber
K Lacy, Jason De Cristofaro, and Jascha Polet
224 The Land-Atmosphere Interactions from
Barometers and Seismometers, Anne Valovcin, Jiong
Wang, and Toshiro Tanimoto
225 Characterizing Seismicity with High-Precision
Relocations of Recent Earthquake Sequences in
Eastern California and Western Nevada, Rachel L
Hatch, Daniel T Trugman, Ken D Smith, Peter M Shearer,
and Rachel E Abercrombie
226 Preliminary report on site characterization using
noninvasive single- and multi-station methods at
southern California seismic stations, Alan Yong,
Antony Martin, Jennifer Pfau, Devin McPhillips, Marcos
Alvarez, Scott S Lydeen, Fiona Clerc, and Nolan Leue
227 Using Waveform Modeling to Determine the Depth
of the Deepest Earthquakes on and Near the
WELCOME
2016 SCEC Annual Meeting page 143
Newport-Inglewood Fault Zone in the Los Angeles
Basin, Chloe S Sutkowski, and Jascha Polet
228 Apparent Attenuation at High Frequencies in
Southern California, Yu-Pin Lin, and Thomas H Jordan
229 Towards automated estimates of directivity and
related source properties of small to moderate
earthquakes in Southern California with second
seismic moments, Haoran Meng, Yehuda Ben-Zion,
and Jeff J McGuire
230 Mento Carlo Inversion of the 1D velocity and
anisotropic model in the Juan de Fuca plate ,
Tian Feng, Thomas Bodin, and Lingsen Meng
231 Exploring the suitability of the Quake-Catcher
Network in the USGS ShakeMap: Case studies from
aftershocks of the 2010-2011 Darfield and
Christchurch, New Zealand earthquakes, Liam J
Shaughnessy, Danielle F Sumy, Elizabeth S Cochran,
Corrie J Neighbors, and Robert-Michael de Groot
232 Differential Waveform Analysis to Test for a Mid-
Crustal Low-Velocity Zone Beneath the Western
Mojave, William K Eymold, Thomas H Jordan, and David
A Okaya
233 Microseismic event detection using local
coherence: applications to the Long Beach 3D array
and the Hi-CLIMB linear array, Zhigang Peng, Zefeng
Li, Dongdong Yao, Daniel D Hollis, and William Frank
234 Evaluation of Ambient Noise Green’s Functions
Using Synthetic Ground Motions, Santiago Rabade,
Leonardo Ramirez-Guzman, Alan Juarez, and Jorge
Aguirre
235 Slip history of the 2016 Mw 7.0 Kumamoto
earthquake: intra-plate rupture in complex tectonic
environment, Chen Ji, Jinglai Hao, and Zhenxing Yao
236 Stochastic Representations of Seismic Anisotropy:
Locally Isotropic Transversely Isotropic Media, Xin
Song, and Thomas H Jordan
237 Status of ShakeAlert and G-FAST in the Pacific
Northwest, John E Vidale, Brendan W Crowell, J R
Hartog, Paul Bodin, Victor C Kress, and Ben Baker
238 Measuring aseismic slip through characteristically
repeating earthquakes at the Mendocino Triple
Junction, Northern California, Kathryn Materna,
Taka'aki Taira, and Roland Bürgmann
239 Remotely triggered small earthquakes along the
San Jacinto Fault Zone, CA, Jennifer M Tarnowski,
and Abhijit Ghosh
240 An energy-based smoothing constraint and the
uncertainty range of co-seismic stress drop of large
earthquakes, Mareike N Adams, Jinglai Hao, Cedric
Twardzik, and Chen Ji
241 Stress Drop and Source Scaling of Recent
Earthquake Sequences across the United States,
Qimin Wu, and Martin C Chapman
242 Tectonic tremor in the San Jacinto Fault, near the
Anza Gap, detected by multiple mini seismic arrays,
Alexandra A Hutchison, and Abhijit Ghosh
243 Strike-slip Faulting Energy Release, Lingling Ye,
Hiroo Kanamori, Thorne Lay, and Jean-Philippe Avouac
244 Using a fast similarity search algorithm to identify
repeating earthquake sequences, Nader Shakibay
Senobari, and Gareth J Funning
245 An Investigation into the Eastern Extent of the
Garlock Fault using Ground-based Magnetics and
Seismotectonic Analysis, Brad Ruddy, and Jascha
Polet
246 Low wave speed zones in the crust beneath SE
Tibet revealed by ambient noise adjoint
tomography, Min Chen, Hui Huang, Huajian Yao, Rob
van der Hilst, and Fenglin Niu
247 Comparisons between the 2016 USGS induced-
seismicity hazard model and “Did You Feel It?” data,
Isabel White, Taojun Liu, Nicolas Luco, and Abbie B Liel
WELCOME
2016 SCEC Annual Meeting page 144
248 Dynamic Triggering and fault geometry in the
Central Himalaya Seismic Gap, Manuel M Mendoza,
Abhijit Ghosh, and S S Rai
249 MyShake: Initial Observations from a Global
Smartphone Seismic Network, Qingkai Kong, Richard
M Allen, and Louis Schreier
251 Strong along-strike variation in aftershock
distribution and rupture propagation of 2015 Mw 7.8
Gorkha earthquake in Nepal, Abhijit Ghosh, Bo LI, and
Manuel M Mendoza
252 Geophysical characterization of twelve CSMIP
station sites in Riverside County, California, Antony
Martin, Lauren Demine, William Dalrymple, Nolan Leue,
and David Carpenter
253 Constraining the velocity structure near the
tremorogenic portion of the San Andreas Fault near
Parkfield, CA using Ambient Noise Tomography,
Rachel C Lippoldt, Robert W Porritt, and Charles G
Sammis
254 Comparison of LASSIE observed travel times and
amplitudes with predicted values from SCEC
velocity models, Zagid Abatchev, and Paul M Davis
255 Aftershock Productivity on Volcanoes: What can it
tell us about interpreting aftershocks? Ricardo Garza-
Giron, Emily E Brodsky, and Stephanie G Prejean
256 Seismogeodetic Observations of the June 10, 2016
M5.2 Borrego Springs Earthquake, Dara E Goldberg,
Jessie K Saunders, Jennifer S Haase, and Yehuda Bock
257 Fault-Zone Exploration in Highly Urbanized Settings
Using Guided Waves: An Example From the
Raymond Fault, Los Angeles, California, Rufus D
Catchings, Janis L Hernandez, Robert R Sickler, Mark R
Goldman, Joanne H Chan, and Coyn J Criley
258 Magnitude and Peak Amplitude Relationship for
Microseismicity Induced by Hydraulic Fracture
Experiment, Trevor Smith, Adam Arce, and Chen Ji
259 Spatial and temporal relationships between tremor
and slip in large slow slip earthquakes on the
Cascadia subduction zone, Heidi Houston, Kelley A
Hall, and David A Schmidt
260 Deformation of the Xishancun Landslide, Sichuan,
inferred from seismicity, Risheng Chu
261 Optical Fiber Strainmeters: Developing Higher
Dynamic Range and Broader Bandwidth, Billy
Hatfield, Mark A Zumberge, Frank K Wyatt, and Duncan
C Agnew
262 Seismic Evidence for Splays of the Eureka Peak
Fault beneath Yucca Valley, California, Mark R
Goldman, Rufus D Catchings, Joanne Chan, Robert R
Sickler, Coyn Criley, Dave O'Leary, and Alan Christensen
263 A derivation of the median ratios between different
definitions of horizontal component of ground
motions in Central and Eastern United States,
Alireza Haji-Soltani, and Shahram Pezeshk
Ground Motion Simulation Validation (GMSV) Posters 281-297
281 Incorporation of Local Site Effects in Broadband
Simulations of Ground Motions: Case Study of the
Wildlife Liquefaction Array, Ramin Motamed, and John
G Anderson
282 Site Response During And After Nonlinear Soil
Behavior Has Occurred, Kenneth S Hudson
283 Site amplification effects at Heathcote Valley, New
Zealand, during the 2010-2011 Canterbury
earthquakes, Seokho Jeong, and Brendon A Bradley
284 Effects of realistic fault geometry on simulated
ground motions in the 2010 Darfield Earthquake,
New Zealand, Hoby N Razafindrakoto, Brendon A
Bradley, and Robert W Graves
WELCOME
2016 SCEC Annual Meeting page 145
285 Hybrid Broadband Ground motion simulations of
Porters Pass fault earthquakes, Mohammad A Nazer,
Hoby N Razafindrakoto, and Brendon A Bradley
286 Ground Motion Simulation Validation using Small-
to-Moderate Magnitude Events in the Canterbury,
New Zealand Region, Robin L Lee, Brendon A Bradley,
and Seokho Jeong
287 Near source ground-motion modelling of the
Canterbury aftershocks and implications for
engineering metrics, Anna E Kaiser, and Caroline
Holden
288 Validation of 3D velocity models using earthquakes
with shallow slip: case study of the Mw6.0 2014
South Napa, California, event, Walter Imperatori, and
Frantisek Gallovic
289 Simulation of multi-segment kinematic rupture of the
1992 Landers earthquake and ground motion
validation of ground motion, Jorge G Crempien, and
Ralph J Archuleta
290 Validation of ground-motion simulations using
precarious rocks, Elliot K Bowie, Mark W Stirling, and
Chris Van Houtte
291 Toppling of PBRs Exposed to Ground Motions
Estimated from the Composite Source Model of
Earthquakes, Richard J Brune, John G Anderson, Glenn
P Biasi, and James N Brune
292 Inter-Period Correlations of SCEC Broadband
Platform Fourier Amplitudes and Response
Spectra, Jeff R Bayless, Paul G Somerville, Andreas
Skarlatoudis, Fabio Silva, and Norman A Abrahamson
293 Kinematic source generation based on fractal fault
geometries, William H Savran, and Kim B Olsen
294 Surface topography effects in three-dimensional
physics-based deterministic ground motion
simulations in southern California, Andrea C Riaño,
Doriam Restrepo, Ricardo Taborda, and Jacobo Bielak
295 Topography effects in the near field for a dislocation
point source: A comparison of the FEM-HERCULES
code and the 3D IBEM, Edilson F Salazar Monroy,
Leonardo Ramirez-Guzman, Marcial Contreras-Zazueta,
and Francisco J Sanchez-Sesma
296 Guidance on the utilization of ground motion
simulations in engineering practice, Brendon A
Bradley, Didier Pettinga, and Jack W Baker
297 Performance-Based Fault Displacement Estimation
With Uniform California Earthquake Rupture
Forecast 3 Probabilities, Glenn P Biasi
Working Group on California Earthquake Probabilities (WGCEP) Posters 298-300
298 A Spatiotemporal Clustering Model for the Third
Uniform California Earthquake Rupture Forecast
(UCERF3-ETAS) – Toward an Operational
Earthquake Forecast, Edward H Field, Kevin R Milner,
Jeanne L Hardebeck, Morgan T Page, Nicholas J van der
Elst, Thomas H Jordan, Andrew J Michael, Bruce E Shaw,
and Maximilian J Werner
299 Testing ETAS Catalogs from UCERF3, Morgan T
Page, Nicholas J van der Elst, Edward H Field, and Kevin
R Milner
300 Trimming the UCERF3-TD Hazard Tree with a New
Probabilistic Model-Reduction Technique, Keith A
Porter, Edward H Field, and Kevin R Milner
Earthquake Forecasting and Predictability (EFP) Posters 301-311 301 Earthquake Number Forecasts Testing, Yan Y Kagan,
and David D Jackson
WELCOME
2016 SCEC Annual Meeting page 146
302 Precursory Swarm and Earthquake Forecasting in
Indo-Nepal Himalaya, Prof Daya SHANKER, and
Harihar Paudyal
303 ETAS 2.x: The 2015 Nepal aftershock forecast, a
Global ETAS Code, and ETAS based Ground
Motion Forecastint, Mark R Yoder, John M Wilson,
John B Rundle, and Margaret T Glasscoe
304 Earthquake likelihood models for New Zealand
combining information on strain rates, earthquakes
and faults, David A Rhoades, Annemarie
Christophersen, and Matthew C Gerstenberger
305 Seismic and Aseismic Moment Budget and
Implication for the Seismic Potential of the Parkield
Segment of the San Andreas Fault, Sylvain G Michel,
Jean-Philippe Avouac, Romain Jolivet, and Lifeng Wang
306 Systematic fluctuations in the global seismic
moment release, Corné W Kreemer, and Ilya Zaliapin
307 Correlation between Strain Rate and Seismicity and
Its Implication for Earthquake Rupture Potential in
California and Nevada, Yuehua Zeng, Mark D
Petersen, and Zheng-Kang Shen
308 Testing the effect of deficient real-time earthquake
catalogs on non-Poissonian earthquake likelihood
models: Examples from the Canterbury earthquake
sequence, Annemarie Christophersen, David A
Rhoades, David S Harte, and Matthew C Gerstenberger
309 Evaluating the use of declustering for induced
seismicity hazard assessment, Andrea L Llenos, and
Andrew J Michael
310 Earthquake Declustering via a Nearest-Neighbor
Approach, Ilya Zaliapin, and Yehuda Ben-Zion
311 Sedimentation in Nearshore Basins as Related to
Paleoseismicity, Will Berelson, Alex Sessions, Josh
West, and James F Dolan
Collaboratory for the Study of Earthquake Predictability (CSEP) Posters 312-315 312 Dependence of b-value on depth, co-seismic slip,
and time for large magnitude earthquakes, John M
Aiken, Takahiko Uchide, and Danijel Schorlemmer
313 Regional evolution of network detection
completeness in Japan, Danijel Schorlemmer, Naoshi
Hirata, Yuzo Ishigaki, Kazuyoshi Z Nanjo, Hiroshi
Tsuruoka, Thomas Beutin, and Fabian Euchner
314 CSEP Evaluations of 24-Hour Earthquake
Forecasting Models for California: New Results and
Ensemble Models, Maximilian J Werner, Maria Liukis,
Warner Marzocchi, David A Rhoades, Matteo Taroni,
Zechar D Zechar, and Thomas H Jordan
Collaboratory for Interseismic Simulation and Modeling (CISM) / Earthquake Simulators Posters 316-322
316 Simulation Based Earthquake Forecasting with
RSQSim, Jacquelyn J Gilchrist, Thomas H Jordan, James
H Dieterich, and Keith B Richards-Dinger
317 Mitigating Induced Seismicity Through Active
Pressure Management: Simulation-Based Studies,
Kayla A Kroll, Keith B Richards-Dinger, and Joshua A
White
318 Using Virtual Quake for GNSS tsunami early
warning: coupling earthquakes, tsunamis, and
ionosphere physics, John M Wilson, Kasey W Schultz,
John B Rundle, and Ramya Bhaskar
319 Fakequakes: Earthquake Rupture Scenarios for
California and Cascadia, Christine J Ruhl, Diego
Melgar, Ronni Grapenthin, Mario Aranha, and Richard M
Allen
WELCOME
2016 SCEC Annual Meeting page 147
320 Provably stable and high-order accurate finite
difference schemes on staggered grids and their
potential application for seismic hazard analysis,
Ossian O'Reilly, and Eric M Dunham
321 Earthquake cycle simulations with rate-and-state
friction and linear and nonlinear viscoelasticity, Kali
L Allison, and Eric M Dunham
Communication, Education, and Outreach (CEO) Posters 323-335 323 Beat the Quake: Designing an Earthquake Themed
Puzzle Room, Jenny Novak, and Julian C Lozos
324 Challenges and Opportunities for Communicating
about Seismic Hazard, Timothy L Sellnow, Emina
Herovic, and Deanna Sellnow
325 www.fallasdechile.cl, the First Online Repository for
Neotectonic Faults in the Chilean Andes, Felipe
Aron, Valeria Salas, Constanza Bugueño, César
Hernández, Luis Leiva, Isabel Santibáñez, and José
Cembrano
326 Music of Earthquakes, Debi Kilb, and Daniel T
Trugman
327 Developing the Next Generation SCEC Web
Infrastructure, Edric Pauk, David Gill, Tran T Huynh,
Mark L Benthien, John E Marquis, Jason Ballmann, John
Yu, and Philip J Maechling
328 SCEC Web in the Clouds: Motivations and
Experiences, David J Gill, Philip J Maechling, Tran T
Huynh, John E Marquis, John Yu, Edric Pauk, Jason
Ballmann, Mark L Benthien, Deborah Gormley, and Annie
Lo
329 2016 SCEC Undergraduate Studies in Earthquake
Information Technology (UseIT): Earthquake
Forecasting Through Physics-Based Simulations,
Zhenyu Fu, Morgan T Bent, Hernan M Lopez, Spencer P
Ortega, and Kevin C Scroggins
331 2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): Earthquake
Magnitude & Event Probability Forecasting, Daniel
Ngu, Brady Guan, Mingyan Zhou, Yipeng Li, Jozi K
Pearson, Thomas H Jordan, Kevin R Milner, and Mark L
Benthien
332 2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): SCEC-VDO
Development & Visualization, Rory C Norman, Kishan
Rajasekhar, Emy S Huang, Patrick Vanderwall, Allen Y Shi,
Rafael G Cervantes, Mark L Benthien, Thomas H Jordan,
FNU Sanskriti, Jozi K Pearson, John Yu, and Kevin R
Milner
333 2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): Loss Analysis of
Earthquake Sequences, Sophia Belvoir, Vianca E
Severino Rivas, Jenepher M Zamora, Jadson da Silva
Lima, Luis J Gomez, Jozi K Pearson, Mark L Benthien, and
Thomas H Jordan
334 2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): Documentary and
Virtual Reality Game, Geneva I Burkhardt, Aadit R
Doshi, Emily R Olmos, Emma L Huibregtse, Drew P
Welch, Mark A Romano, Jason Ballmann, Jozi K Pearson,
and Mark L Benthien
335 SCEC-VDO Reborn: A New 3-Dimensional
Visualization and Movie Making Software for Earth
Science Data, Kevin R Milner, FNU Sanskriti, John Yu,
Scott Callaghan, Philip J Maechling, and Thomas H
Jordan
Community Modeling Environment (CME) / Computational Science (CS) Posters 336-350
WELCOME
2016 SCEC Annual Meeting page 148
336 AWP-ODC Open Source Project, Dawei Mu, Alex
Breuer, Josh Tobin, Hui Zhou, and Yifeng Cui
337 CyberShake Advancements: Broadband
Seismograms and Central California Sites, Scott
Callaghan, Philip J Maechling, Christine A Goulet, Karan
Vahi, Robert W Graves, Kim B Olsen, Kevin R Milner,
David Gill, Yifeng Cui, and Thomas H Jordan
338 Next Generation Boundary Element Models for
Earthquake Science, Thomas B Thompson, and
Brendan J Meade
339 Inferring Southern California Crustal Viscosity
Structure from the SCEC Community Velocity
Model, William Shinevar, Mark Behn, Greg H Hirth, and
Oliver Jagoutz
340 Accelerating AWP-ODC-OS Using Intel Xeon Phi
Processors, Josh Tobin, Alexander N Breuer, Charles
Yount, Alexander Heinecke, and Yifeng Cui
341 High performance earthquake simulations in
viscoelastic media using SeisSol, Carsten Uphoff,
Michael Bader, Sebastian Rettenberger, and Alice-Agnes
Gabriel
342 Recent Advances of the ADER-DG Finite Element
Method for Seismic Simulations, Alexander N Breuer,
Alexander Heinecke, and Yifeng Cui
343 Access to Geodetic Imaging Products through
GeoGateway Analysis, Modeling, and Response
Tools, Andrea Donnellan, Jay W Parker, Robert A
Granat, Michael B Heflin, Marlon E Pierce, Jun Wang,
John B Rundle, and Lisa Grant Ludwig
344 Recent Improvements to AWP-ODC-GPU
supported by Blue Waters PAID Program, Yifeng Cui,
Daniel Roten, Peng Wang, Dawei Mu, Carl Pearson, Kyle
B Withers, William H Savran, Kim B Olsen, Steven M Day,
and Wen-Mei Hwu
345 Using Coupled Geomechanical Modeling to
Investigate Potential Production Induced Seismicity,
Robert L Walker, Susan E Hough, Birendra Jha, Victor C
Tsai, Morgan T Page, and Fred Aminzadeh
346 PH5 for integrating and archiving different data
types, Bruce C Beaudoin, Derick Hess, and Steve
Azevedo
347 Performance enhancements and visualization for
RSQSim earthquake simulator, Dmitry Pekurovsky,
Amit Chourasia, Keith B Richards-Dinger, Bruce E Shaw,
James H Dieterich, and Yifeng Cui
348 SeedMe: Data sharing building blocks, Amit
Chourasia, David R Nadeau, John Moreland, Dmitry
Mishin, and Michael L Norman
349 Hybrid Genetic Deflated Newton Method for
Distributed-Source Optimization, Marcus M Noack,
and Steven M Day
350 QuakeCoRE ground motion simulation
computational workflow, Sung Bae, Viktor Polak,
Richard Clare, Brendon A Bradley, and Hoby N
Razafindrakoto
WELCOME
2016 SCEC Annual Meeting page 149
Meeting Abstracts
Flexible implementation of multiphysics and
discretizations in PyLith crustal deformation modeling
software, Brad T Aagaard, Charles A Williams, and Matthew G
Knepley (Poster Presentation 043)
We are creating a flexible implementation of multiphysics and
finite-element discretizations in PyLith, a community, open-
source code (http://geodynamics.org/cig/software/pylith/) for
modeling quasi-static and dynamic crustal deformation with an
emphasis on earthquake faulting. The goals include expanding
the current suite of elastic, viscoelastic, and elastoplastic bulk
rheologies to include poroelasticity, thermoelasticity, and
incompressible elasticity. We cast the governing equations in a
form that involves the product of the finite-element basis function
or its derivatives with pointwise functions that look very much like
the strong form of the governing equation. This allows the finite-
element integration to be decomposed into a routine for the
numerical integration over cells and boundaries of the finite-
element mesh and simple routines implementing the physics
(pointwise functions). The finite-element integration routine
works in any spatial dimension with an arbitrary number of
physical fields (e.g., displacement, temperature, and fluid
pressure). It also makes it much easier to optimize the finite-
element integrations for proper vectorization, tiling, and other
traversal optimization on multiple architectures (e.g., CUDA and
OpenCL) independent of the pointwise functions. Users can
easily extend the code by adding new routines for the pointwise
functions to implement different rheologies and/or governing
equations. Tight integration with the Portable, Extensible Toolkit
for Scientific Computation (PETSc) provides support for a wide
range of linear and nonlinear solvers and time-stepping
algorithms so that a wide variety of governing equations can be
solved efficiently.
Comparison of LASSIE observed travel times and
amplitudes with predicted values from SCEC velocity
models., Zagid Abatchev, and Paul M Davis (Poster
Presentation 254)
A dense broadband survey (LASSIE Los Angeles Seismic
Syncline Interferometry Experiment) was conducted to improve
our velocity and structure models for the Los Angeles Basin
region. The survey was a cooperative experiment among
industry, academia and government and consisted of a total of
73 broadband sensors installed across the LA basin between
Long Beach and the Puente Hills. The data has been shared
amongst all participants. In addition to noise correlation studies,
local earthquakes and teleseisms were well recorded. The data
will be used in a joint inversion for structure. We are computing
travel times and amplitudes for events using our implementation
of a ray shooting model through the 3D velocity structures given
by the SCEC Community Velocity Model (CVM and CVM-H) and
models being derived from LASSIE data. We are comparing
predicted and observed values in order to determine corrections
to the CVM.
Earthquake source parameter validation using multiple-
scale approaches for induced seismicity in Oklahoma,
Rachel E Abercrombie, and Xiaowei Chen (Poster Presentation
063)
The rapid increased seismicity rate in central US has drawn
significant attention to the associated earthquake hazards in
intraplate regions. Whether induced earthquakes have similar
stress drops to tectonic events is an important question for
understanding both the earthquake rupture process, and the
hazard they represent. To address this question, we study well-
recorded earthquake sequences in Oklahoma, including clusters
near both Guthrie, and Prague.
For each cluster, we apply both a stacking-based empirical
Green’s function method and an individual-pair based method to
both P and S waves. Significant effort has been devoted to
improve the stability of the stacking approach. We carefully
examine variations of the average stress drop for each cluster
with different choices of signal-to-noise ratio, minimum
requirement of stations, number of events, magnitude bins
included in stacking, and frequency band limits, in order to
assess a range of parameter choices that provide stable
estimations. In the stacking approach, the average stress drop
would shift the individual stress drop systematically. Although this
would generally preserve the relative variations, this could be
problematic when comparing sequences from different regions.
We then compare the optimal stacking result with individual-pair
analysis for common events. For the latter method, we also
investigate how to obtain and identify the best results using
multiple EGFs and stations. For the Guthrie clusters, we obtain
striking consistency over a wide range of magnitudes, and
consistent average stress drop. Some parameter combinations
affect the average stress drop, which may appear anomalously
low or high. We continue to analyze other sequences to optimize
the algorithms. The stress drop results will help us to understand
the variations of fault strength and relationship with pore pressure
propagation.
Towards Detailed Characterization of Spatio-temporal
Variations in Stress Parameters along the San Jacinto
Fault Zone, Niloufar Abolfathian, Patricia Martínez-Garzón, and
Yehuda Ben-Zion (Poster Presentation 013)
Accurate determination of stress parameters (orientation of
principal stress axes and stress ratio) operating on fault zones
provides refined knowledge on source physics and deformation
processes. We developed a refined stress inversion
methodology using declustered seismicity, to remove events
likely affected by local stress interactions rather than reflecting
the large-scale background stress field, and discretizing the
remaining focal mechanisms using an updated k-means
technique [Martínez-Garzón et al., in prep].
In this study we apply the refined stress inversion methodology
to obtain a reliable high resolution information of stress
parameters throughout the San Jacinto Fault Zone in Southern
California. In particular, we investigate potential stress changes
along and across the fault and also through the seismogenic
depth. The stress parameters are analyzed using two seismicity
catalogs containing focal mechanisms before and after the 2010,
WELCOME
2016 SCEC Annual Meeting page 150
Mw 7.2 El Mayor-Cucapah earthquake in Baja California, to
detect potential stress changes in relation to the occurrence of
this event.
The initial results are in agreement with the expected stress type
due to the tectonic setting and the geology of the region. Both
inversions before and after El Mayor earthquake are generally
consistent along the fault, showing transtensional stress regime
in the northwestern section and more complex oblique faulting
within the main fault traces in the middle and southeastern
sections. These results are in agreement with studies by Yang
and Hauksson (2013) and Bailey et al. (2010). Changes of stress
shape ratio across the fault are observed in selected regions
such as the trifurcation area, while the orientation of the
maximum horizontal stress is consistent across the fault. The
continuing work focuses on potential spatial changes in the
stress field related to damage zones and depth, and temporal
variations of stress parameters.
An energy-based smoothing constraint and the
uncertainty range of co-seismic stress drop of large
earthquakes, Mareike N Adams, Jinglai Hao, Cedric Twardzik,
and Chen Ji (Poster Presentation 240)
The large uncertainties of some basic earthquake source
parameters, such as average stress drop (Cotton et al., 2013),
are largely caused by over-simplifications of the source (e.g.,
Madariaga, 1979). In contrast, the results of finite fault source
inversions suffer from the uncertainties caused by over-
parameterization. Hundreds of free parameters are often used to
represent spatiotemporal rupture history, despite the fact that not
all of them are well constrained by the available observations.
The uncertainties of individual sub-fault parameters are often
difficult to access because they are not only caused by the limited
data but also the regularizations applied to stabilize the
inversions. We develop a new finite fault inversion strategy to
explore the uncertainty range for the energy based average
stress drop (Δτ_E ) of large earthquakes. For a given earthquake,
a series of modified finite fault inversions are conducted to search
for the solution that not only best fits seismic and geodetic data
but also has a Δτ_E matching a given value. It results in a trade-
off curve between the misfit to the observations and Δτ_E,
allowing the range of Δτ_E constrained by the given geophysical
dataset to be robustly defined. Our initial study of the 2014 Mw7.9
Rat Islands earthquake using teleseismic data revealed that only
the lower bound of Δτ_E could be constrained with far field
seismic data (Adams et al., 2016). To investigate whether such
a conclusion also holds for near field data, the slip distribution
and Δτ_E of the 2015 Mw7.9 Gorkha, Nepal earthquake is
studied using near-source geodetic data. Our result reveals a
similar pattern that the misfit gradually improves when the target
Δτ_E increases from 0.5 to 8 MPa; but becomes nearly constant
from 10 MPa to 50 MPa. Hence, only the lower bound (~8-10
MPa) of Δτ_E can be constrained, even with such a
comprehensive geodetic dataset. As Δτ_E is proportional to the
available seismic energy, our results imply that only the lower
bound of the available energy can be constrained. We notice that
the Laplacian of fault slip monotonously increases with Δτ_E and
propose to use Δτ_E as a new regularization tool to smooth fault
slip during the source inversion. The inverted result can be
viewed as the solution with minimum available seismic energy for
one earthquake.
Dependence of b-value on depth, co-seismic slip, and
time for large magnitude earthquakes, John M Aiken,
Takahiko Uchide, and Danijel Schorlemmer (Poster Presentation
312)
Spatial and temporal variations in parameters of earthquake
sources and seismicity will be a key to understand the status of
faults such as applied stress and fault strength and hence the
potential of earthquake occurrence. Recent studies have shown
that the b-value of the Gutenberg-Richter relationship decreases
before large magnitude events for timescales of five years or
more. In addition, large co-seismic slip values have been found
to occur in extremely low b-value regions, e.g., the 2011 Tohoku-
oki earthquake [Nanjo, et. al. 2012]. A comprehensive
understanding requires us to explore the relationship between
depth, co-seismic slip of a large earthquake, time, and b-value
for large magnitude events occurring in different tectonic
environments (oceanic and in-land). Here we present work
studying the relationship between co-seismic slip and b-values
prior to two large magnitude events in Japan, the 2016 M7.0
Kumamoto earthquake and the 2011 Tohoku-oki earthquake,
using the Japan Meteorological Agency (JMA) earthquake
catalog. We calculated the b-values at three-dimensionally
distributed grids using the maximum likelihood estimation
method and the reported JMA magnitude of completeness. We
examined the cross section of depth, co-seismic slip, and time
with the calculated b-values to investigate whether a common
behavior exists prior to large events. These results will be used
in designing a final model submitted to the Collaboratory for the
Study of Earthquake Predictability (CSEP), Japan for testing.
Constraints on Fault Damage Zone Properties and
Normal Modes from a Dense Linear Array Deployment
along the San Jacinto Fault Zone, Amir A Allam, Fan-Chi Lin,
Pieter-Ewald Share, Yehuda Ben-Zion, Frank L Vernon, Gerard
Schuster, and Marianne Karplus (Poster Presentation 201)
We present earthquake data and statistical analyses from a
month-long deployment of a linear array of 134 Fairfield three-
component 5 Hz seismometers along the Clark strand of the San
Jacinto fault zone in Southern California. With a total aperture of
2.4km and mean station spacing of ~20m, the array locally spans
the entire fault zone from the most intensely fractured core to
relatively undamaged host rock on the outer edges. We recorded
36 days of continuous seismic data at 1000Hz sampling rate,
capturing waveforms from 751 local events with Mw>0.5 and 43
teleseismic events with M>5.5, including two 600km deep M7.5
events along the Andean subduction zone. For any single local
event on the San Jacinto fault, the central stations of the array
recorded both higher amplitude and longer duration waveforms,
which we interpret as the result of damage-related low-velocity
structure acting as a broad waveguide. Using 271 San Jacinto
events, we compute the distributions of three quantities for each
station: maximum amplitude, mean amplitude, and total energy
(the integral of the envelope). All three values become
statistically lower with increasing distance from the fault, but in
addition show a nonrandom zigzag pattern which we interpret as
normal mode oscillations. This interpretation is supported by
polarization analysis which demonstrates that the high-amplitude
late-arriving energy is strongly vertically polarized in the central
part of the array, consistent with Love-type trapped waves. These
results, comprising nearly 30,000 separate coseismic
WELCOME
2016 SCEC Annual Meeting page 151
waveforms, support the consistent interpretation of a 450m wide
asymmetric damage zone, with the lowest velocities offset to the
northeast of the mapped surface trace by 100m. This asymmetric
damage zone has important implications for the earthquake
dynamics of the San Jacinto and especially its ability to generate
damaging multi-segment ruptures.
Earthquake cycle simulations with rate-and-state friction
and linear and nonlinear viscoelasticity, Kali L Allison, and
Eric M Dunham (Poster Presentation 321)
We have implemented a parallel code that simultaneously
models both rate-and-state friction on a strike-slip fault and off-
fault viscoelastic deformation throughout the earthquake cycle in
2D. Because we allow fault slip to evolve with a rate-and-state
friction law and do not impose the depth of the brittle-to-ductile
transition, we are able to address: the physical processes limiting
the depth of large ruptures (with hazard implications); the degree
of strain localization with depth; the relative partitioning of fault
slip and viscous deformation in the brittle-to-ductile transition
zone; and the relative contributions of afterslip and viscous flow
to postseismic surface deformation.
The method uses a discretization that accommodates variable
off-fault material properties, depth-dependent frictional
properties, and linear and nonlinear viscoelastic rheologies. All
phases of the earthquake cycle are modeled, allowing the model
to spontaneously generate earthquakes, and to capture afterslip
and postseismic viscous flow. We compare the effects of a linear
Maxwell rheology, often used in geodetic models, with those of a
nonlinear power law rheology, which laboratory data indicates
more accurately represents the lower crust and upper mantle.
The viscosity of the Maxwell rheology is set by power law
rheological parameters with an assumed a geotherm and strain
rate, producing a viscosity that exponentially decays with depth
and is constant in time. In contrast, the power law rheology will
evolve an effective viscosity that is a function of the temperature
profile and the stress state, and therefore varies both spatially
and temporally. We will also integrate the energy equation for the
thermomechanical problem, capturing frictional heat generation
on the fault and off-fault viscous shear heating, and allowing
these in turn to alter the effective viscosity.
Towards Implementation of Multi-Segment Ruptures in
the Broadband Platform: Composite Source Model, John
G Anderson (Poster Presentation 265)
One major goal of the Broadband Platform for 2016 is to
implement multi-segment ruptures. One straightforward
approach under consideration is to generate ground motions
from each segment separately, and then add them with an
appropriate time lag. Research on this is still underway, but some
preliminary results can be reported on two issues: scaling of the
two rupture segments, and sensitivity to the approach described
above.
The scaling issue is illustrated by considering two segments of
the San Jacinto fault considered by Lozos et al. (2015): the
Claremont strand (L1=79 km) and the Casa Loma – Clark strand
(L2=110 km). With the 21 km overlap of the strands, the end-to-
end length is L3=168 km. For modeling, we need the sum of the
moments of the two strands to equal the moment of the entire
fault (Mo(L1)+Mo(L2)=Mo(L3)). Defining
r=(Mo(L1)+Mo(L2))/Mo(L3), and using Wells and Coppersmith
(1994) or Hanks and Bakun (2008), r= 0.77 and 0.65,
respectively. Recently, we (Anderson, Biasi and Wesnousky,
2016) have proposed a scaling model that aims to have a
constant stress drop over the entire range rupture lengths. This
model gives r=1.03 for the above case, which is closer to unity
than the other considered models, and thus an improvement.
However, the ABW16 model is not so helpful when there are
short fault segments, so additional thought is still required.
Examples of the effects of summing ground motions from the
different segments, will be shown for the San Jacinto fault,
emphasizing the cases of precariously balanced rocks discussed
in the poster by Brune et al. nearby at this meeting.
Quaternary Expression of Northern Great Valley Faults
and Folds: Accommodating North-South Contraction in
the Northeastern California Shear Zone, Stephen J Angster,
Thomas Sawyer, and Steven G Wesnousky (Poster Presentation
094)
The Northern California Shear Zone accommodates North
American intraplate right-lateral transpressional shear driven by
the relative motions of the northwest translating Sierran
microplate and clockwise rotating Oregon Costal block. Within
this zone, between the latitudes of 40 and 42, 1 – 4 mm/yr of
north-south oriented contraction is geodetically observed and for
the most part remains unaccounted for in the geology. The
northeast trending Inks Creek fold belt north of Red Bluff, CA
consists of anticline-syncline pairs and a dome structure that
appear optimally oriented to accommodate northwest crustal
shortening. The anticlines plunge to the southwest where they
are incised by and appear to deflect the course of the
Sacramento River. Geomorphic mapping and profiling of fluvial
terraces along the Sacramento River are interpreted to record
tectonic uplift through the fold belt. Dating of elevated fluvial
terraces on the anticlines with radiocarbon, terrestrial
cosmogenic nuclide, and OSL dating techniques is currently
underway to quantify the vertical incision rate of the river as a
result of tectonic uplift. The ultimate goal is to combine these
geomorphic observations with structural observations and
models to estimate the rate of contraction being accommodated
by the folds.
Investigating the high-frequency, weak-motion nucleation
phase initiating the June 10th, 2016 Mw 5.2 Borrego
Springs Earthquake, Adam Arce, Mareike N Adams, and Chen
Ji (Poster Presentation 057)
A nucleation phase, when present, offers insight into an
earthquake’s rupture mechanics and may corroborate the
“cascade” rupture model. In cascade rupture, slip initiates on a
relatively small patch progressing into slip on a larger patch and
so on, as accumulated strain continues to exceed the frictional
stress along the fault. The nucleation phase is recorded as weak
ground motion preceding the strong ground motion of a given
event and has been observed with earthquakes of moment
magnitude (Mw) ranging 2.1 – 8.1 (Beroza and Ellsworth, 1996).
On June 10th 2016 a Mw 5.2 earthquake occurred 20 kilometers
NNW of Borrego Springs with moderate shaking reported
(USGS) more than 80 km away. Strong motion data from 15
stations, ranging ~15 – 50 km from the source with good
WELCOME
2016 SCEC Annual Meeting page 152
azimuthal coverage, recorded a high-frequency, weak-motion
signal during the 1.70 – 1.76 seconds preceding the arrival of
strong-motion phases. Ellsworth and Beroza (1995) observed a
scaling relationship between nucleation duration and seismic
moment, likely dependent on a constant stress-drop. The scaling
law predicts a nucleation duration of ~ 0.12 s for the Borrego
Springs event, an order of magnitude smaller than what is
observed. A finite fault source inversion is being performed on
the velocity seismograms of the strong-motion data. Further
investigation is being done to understand the anomalous
duration and what role this nucleation phase may have played in
rupture of the Borrego Springs event.
www.fallasdechile.cl, the First Online Repository for
Neotectonic Faults in the Chilean Andes, Felipe Aron,
Valeria Salas, Constanza Bugueño, César Hernández, Luis
Leiva, Isabel Santibáñez, and José Cembrano (Poster
Presentation 325)
We introduce the site www.fallasdechile.cl, created and
maintained by undergraduate students and researchers at the
Catholic University of Chile. Though the web page seeks to
inform and educate the general public about potentially
seismogenic faults of the country, layers of increasing content
complexity allow students, researchers and educators to consult
the site as a scientific tool as well. This is the first comprehensive,
open access database on Chilean geologic faults; we envision
that it may grow organically with contributions from peer
scientists, resembling the SCEC community fault model for
southern California. Our website aims at filling a gap between
science and society providing users the opportunity to get
involved by self-driven learning through interactive education
modules.
The main page highlights recent developments and open
questions in Chilean earthquake science. Front pages show first
level information of general concepts in earthquake topics such
as tectonic settings, definition of geologic faults, and space-time
constraints of faults. Users can navigate interactive modules to
explore, with real data, different earthquake scenarios and
compute values of seismic moment and magnitude. A second
level covers Chilean/Andean faults classified according to their
geographic location containing at least one of the following
parameters: mapped trace, 3D geometry, sense of slip,
recurrence times and date of last event. Fault traces are
displayed on an interactive map using a Google Maps API. The
material is compiled and curated in an effort to present, up to our
knowledge, accurate and up to date information. If interested, the
user can navigate to a third layer containing more advanced
technical details including primary sources of the data, a brief
structural description, published scientific articles, and links to
other online content complementing our site. Also,
geographically referenced fault traces with attributes (kml,
shapefiles) and fault 3D surfaces (contours, tsurf files) will be
available to download.
Given its potential for becoming a referential database for active
faults in Chile, this project evidences that undergrads can go
beyond the classroom, be of service to the scientific community,
and make contributions with broader impacts.
From VS30 to near-surface velocity profiles: Integrating
soft sediments in SCEC's UCVM, Domniki Asimaki, Jian Shi,
and Ricardo Taborda (Poster Presentation 266)
The near-surface soil layers of sedimentary basins play a critical
role in modifying the amplitude, frequency and duration of
earthquake ground shaking. These phenomena, referred to as
site effects, play a very important role in ground-motion
simulations, in the development of site amplification factors, and
more generally in earthquake hazard and risk predictions on a
regional scale. In this study, we develop a model that translates
Vs30, the only proxy available to describe the stiffness of the
near surface sediments in Southern California, into a generic
velocity profile suitable for use in wave propagation-based
ground motion models. We specifically develop a Vs30-
dependent shear wave velocity model (previously referred to as
Geotechnical Layer or GTL), based on the statistics of a few
hundred measured velocity profiles with VS30 ranging from 150
m/s to 800 m/s. We validate the model by comparing the site
amplification of the measured and the developed VS30-
dependent velocity profiles. We lastly develop and demonstrate
the implementation of a spatially correlated random realization
algorithm, intended to populate the near surface of the 3D UCVM
domain with our VS30-dependent velocity profiles. The next step
of this work is to use the profiles to improve high-frequency
predictions of 3D physics-based ground motion simulations; and
to develop VS30-dependent amplification factors for
implementation in the SCEC broadband platform.
Imaging fault scarps and fault zone evolution near an
oceanic transform fault using high-resolution bathymetry,
Curtis W Baden, George E Hilley, Samuel Johnstone, Robert M
Sare, Felipe Aron, Holly Young, Christopher M Castillo, Lauren
Shumaker, Johanna M Nevitt, Tim McHargue, and Charles Paull
(Poster Presentation 111)
Oceanic transform faults play a fundamental role in plate
tectonics by linking spreading ridge segments to each other.
While ubiquitous, they produce far fewer large earthquakes than
faults along other tectonic boundaries. We use high-resolution (1
m) deep-water bathymetry to examine the structure of, and offset
along, transform faults in the Gulf of California. Collected by an
autonomous underwater vehicle (AUV) by the Monterey Bay
Aquarium Research Institute in March 2012, bathymetry and
shallow seismic data provide detailed observations of fault-zone
geomorphology of an active transform fault hosted in an area
transitioning from continental to oceanic crust.
We found that these transform faults consist of an array of fault
strands that span the observational extent of the AUV
deployment, and whose activity varies in space and time to
produce a complex fault zone not unlike onshore, seismogenic
strike-slip faults. Fault scarps and scarp-like features were
mapped and categorized by degree of lateral continuity and
consistency in trend. Laterally continuous features trending
parallel to the overall transform fault zone exhibit higher
amplitudes and younger morphologic ages, as measured by
diffusion modeling of the profile curvature of mapped features.
We observed a series of small submarine fans that were offset
along the main, apparently active transform fault strand by 10-25
m. The large and consistent offsets of these synchronously
formed features may be explained by seismically induced mass
WELCOME
2016 SCEC Annual Meeting page 153
wasting prior to offset by either smaller-magnitude earthquakes
or creep. Up to two sets of nested fan deposits of similar volume
appear in CHIRP profiles collected by a subbottom profiler
attached to the AUV, suggesting that fans formed periodically in
response to destabilization. If correct, these inferences imply that
oceanic transform faults may display both creep and stick-slip
behavior at various times, which is consistent with their globally
observed low seismic moment release rate.
QuakeCoRE ground motion simulation computational
workflow, Sung Bae, Viktor Polak, Richard Clare, Brendon A
Bradley, and Hoby N Razafindrakoto (Poster Presentation 350)
We present the ground motion simulation computational
workflow that has been developed and is currently in use by
QuakeCoRE researchers. The amount of data and complexity of
computation make the large-scale ground motion simulations
practically impossible to run on a researcher's workstation.
QuakeCoRE started collaboration with New Zealand eScience
Infrastructure (NeSI), the national high performance computing
(HPC) provider, to gain the necessary computational capacity
and execution speed.
The ground motion simulation code runs massively in parallel
taking three key inputs (rupture model, velocity model and station
information) that are constantly refined through many runs
followed by extensive post-processing analysis.
The complete workflow has been recently successfully migrated
to NIWA's Fitzroy cluster and QuakeCoRE has secured 1.2
million core hours that will be essential for large-scale
simulations, such as the Alpine Fault rupture scenarios.
We simplified the workflow making it easier for new users to
learn, and minimized the need for user interaction by automating
the process pipeline. Augmenting the computing power from
HPC, this streamlined workflow significantly improves simulation
productivity.
We plan to develop a new automated process to enable a near-
real time earthquake simulation upon an event of a major
earthquake.
Reconstruction Modeling of Topography and Lithosphere
Dynamics using Western U.S. Strain History within the
Pacific-North America Plate Boundary Zone, Alireza
Bahadori, Bill E Holt, Lucy Flesch, Lijun Liu, Troy Rasbury, Gavin
Piccione, and Rubin Smith (Poster Presentation 008)
The complex deformation history of the western U.S. since 36
Ma involved a dramatic transition from a subduction-dominated
to a transform-dominated margin, with widespread extension
within the interior Basin and Range region. This deformation
history altered the topography, drainage patterns and basins
throughout the southwest. To reconstruct the topography and
lithosphere dynamics we perform a comprehensive analysis of
the plate boundary zone in the western U.S. since 36 Ma. Our
goal is to understand the link between mantle dynamics, crustal
deformation history, and topography evolution. Using position
estimates from McQuarrie and Wernicke [2005], we determine
lithospheric strain rates through time. Using present-day crustal
thickness estimates for initial conditions, and assuming volume
conservation, we integrate strain rates to determine estimates of
crustal thickness changes over time. The topography and crustal
thickness values provide estimates for gradients of gravity
potential energy (GPE), or the body force distribution, within the
lithosphere through time. We next solve the force balance
equations for vertically averaged deviatoric stresses associated
with the body force distribution. Our goal is to investigate the role
of boundary conditions provided by plate interaction and mantle
convection effects. We start by solving for stress boundary
conditions, such that when boundary condition effects are added
to the contribution from GPE distributions, we obtain a total stress
field in accord with the kinematic solution for a given time. We
also plan to incorporate the influences of mantle flow using
tomography-constrained convection simulations that have been
moved backward in time using adjoint methods. These
convection simulations provide the lower boundary conditions in
the dynamic models. Our effort on the dynamics will be combined
with other independent geochemical constraints. Today we see
a strong correlation between fluorine concentrations and helium
isotopes in geothermal springs and wells. Elevated
concentrations are found within regions of rapid transtensional
strain rates within the U.S. Great Basin. We infer that these
elevated concentrations of fluorine reflect a mantle fluid
contribution. We plan to use U-Pb dating of fluorite deposits
within fault-associated veins to provide information on the timing
of mantle fluid input.
Mitigating the spatial biases of back-projection imaging
using a “station-based” slowness calibration, Han Bao,
Lingsen Meng, and Zhipeng Liu (Poster Presentation 208)
The back projection (BP) approach has been widely used for
earthquake source analysis, especially for imaging the
complexity of the dynamic rupture of recent megathrust
earthquakes. However, in the previous studies of several large
earthquakes, such as 2004 M9.0 Sumatra and 2015 M7.8 Nepal
earthquakes, the propagation directions and speeds differ
significantly between the BPs of arrays located in different
continents. This discrepancy indicates that the conventional BP
approach suffers from spatial bias in source locations. In order to
improve the accuracy of BP imaging, a new aftershock-based
correction was introduced into the BP approach to account for
uncertainty of the ray parameter (slowness) caused by
approximating the 3D velocity structures with a 1D reference
model. Such calibrations rely on calculating the spatial
derivatives of travel times between aftershocks. However, the
inferred slowness may inherit additional errors from the
uncertainty of aftershock locations and origin times in routine
earthquake catalogs. To circumvent this issue, here we present
an alternative approach to compute the slowness anomalies.
Instead of computing the spatial derivatives of travel times on the
source side, we calculate the travel-time derivatives between
nearby stations along a similar ray-path. We seek to test such
approximations by imaging various synthetic rupture scenarios
based on travel times predicted by a 3D global P-wave velocity
model (LLNL-G3D). We find that the BPs calibrated by the
“station-based” slowness correction achieve similar
performances in mitigating the spatial biases seen in un-
calibrated BPs. We therefore consider the “station-based”
slowness calibration approach an important addition of the back-
projection analysis.
WELCOME
2016 SCEC Annual Meeting page 154
InSAR MSBAS Time-Series Analysis of Induced
Seismicity in Cushing, Oklahoma, Magali Barba, and Kristy
F Tiampo (Poster Presentation 164)
Since 2009, the number of earthquakes in the central and
eastern United States has dramatically increased from an
average of 24 M ≥ 3 earthquakes a year (1973-2008) to an
average of 193 M ≥ 3 earthquakes a year (2009-2014) (Ellsworth,
2013). Wastewater injection, the deep disposal of fluids, is
considered to be the primary reason for this increase in seismicity
rate (Weingarten et al., 2015). We use Interferometric Synthetic
Aperture Radar (InSAR) to study injection induced seismicity in
Cushing, OK. InSAR data complements seismic data by
providing insight into the surface deformation potentially
correlated with earthquake activity. To study the ground
deformation associated with the induced seismicity and injection
well activity, we develop full-resolution interferograms using raw
radar data from Radarsat-1/2, ALOS, and Sentinel-1. We pair the
SAR images using the small perpendicular baseline approach
(Berardino et al., 2002) to minimize spatial decorrelation. The
paired SAR images are processed into interferograms using the
JPL ISCE software (Gurrola et al., 2010). Using the MSBAS
algorithm (Samsonov et al., 2013, Samsonov and d'Oreye, 2012)
and the JPL GIAnT software (Agram et al., 2013), we construct
a time-series of the cumulative surface displacement integrating
all interferograms for the region.
To correlate the relationship between surface deformation and
wastewater injection, we compare the well locations, depths, and
injection rates with the spatial and temporal signature of the
surface deformation before and after induced earthquakes, filling
in the spatiotemporal gap lacking from seismicity. By monitoring
the surface deformation for wells associated with past and
current induced seismicity, we can implement measures to
mitigate induced seismicity and its social and economic impact.
Dynamic Weakening of Sliding Friction and the Influence
of Gouge Development, Monica R Barbery, Frederick M
Chester, Judith S Chester, and Omid Saber (Poster Presentation
032)
Previous experiments demonstrate that as sliding velocity
approaches seismic slip rates, the coefficient of sliding friction (u)
typically reduces significantly and may reflect flash heating of
surface contacts. Experiments further demonstrate differences in
the weakening behavior of bare rock and gouge-lined surfaces
across different rock types. We conducted quasi-static (QS)
experiments and high-speed velocity-step (VS) experiments on
ground surfaces of granite (Westerly) and quartzite (Sioux) using
a double direct shear (DDS) configuration, with sliding area of
75cm2, to investigate the effects of gouge generation and
accumulation on the frictional weakening behavior. High-speed
infra-red imagery was used to measure sliding surface
temperatures. QS experiments were conducted at 6-8 MPa
normal stress and with sliding velocities of 1 mm/s for 36 mm of
displacement. VS experiments were conducted at 7-9 MPa
normal stress and achieved steps from 1 mm/s to 1 m/s at high
acceleration (100g) over a small displacement (2 mm), and with
sustained high-speed sliding for 30 mm. Successive experiments
were conducted without disassembling the blocks or disturbing
the sliding surfaces to generate and accumulate gouge for
cumulative displacements up to 0.5 m. For granite at QS rates, u
strengthened from ~0.75 to 0.8 within ~10 mm of displacement.
For quartzite, u weakened from ~0.65 to 0.6 within a typical
weakening distance, dc, of ~3 mm. In VS tests, locally high
temperatures were observed correlating to corrugated structures
within the gouge. For VS granite tests, an abrupt decrease in u
from ~0.7 at quasi static slip rates to 0.5 at m/s slip rates was
observed over a dc of ~3 mm, consistent with rotary shear
experiments conducted at similar conditions. With the
accumulation of gouge on the sliding surface, dc progressively
increases to ~2 cm. In contrast, VS tests on quartzite show an
abrupt increase in u, from ~0.65 to 0.7 within 1 mm of slip,
followed by gradual weakening for the duration of high-speed
sliding. As quartz gouge accumulated on the sliding surface,
similar behavior is observed, but with a slightly greater magnitude
of strengthening. The quartzite results are unlike those reported
in the literature for quartzite at similar conditions. Gouge is
generated at a greater rate in quartzite when compared to the
granite, and is particularly abundant during the initial stage of slip.
The different behavior of quartzite is currently under further
investigation.
Ground Strains in Southern California from Earthquakes
on the San Jacinto Fault, Andrew J Barbour, and Annemarie
S Baltay (Poster Presentation 278)
Strainmeters record high-frequency seismic waves with signal-
to-noise ratios that can be comparable to those of traditional
seismometers, for average phase dispersion characteristics in
Southern California. Yet, to date, the efficacy of using
strainmeters for ground motion studies has not been
investigated. Here we compare direct and inferred strains from
all M ≥ 4.5 earthquakes on the San Jacinto Fault since the Plate
Boundary Observatory borehole network was installed (5
events), finding that measured strains are comparable to
estimates from PGV (peak ground velocity) scaled by S-wave
slowness. We compare the strains and converted PGV values for
these events to NGA-West2 ground-motion prediction equations
(GMPEs), also scaled by the S-wave slowness. While there is an
overall match between the data and GMPEs, we observe a
significant departure from expected attenuation behavior: peak
inferred strains in the Los Angeles basin resulting from moderate
earthquakes in the Anza region are significantly larger than
predicted by GMPEs, and cannot be understood in terms of site-
to-site variations in Vs30 alone. We thus explore the effects of
basin amplification in the GMPEs. Together, the strain and
seismic network infrastructure in California are complementary
networks that can be used for hazard assessment in critical
areas.
High-Resolution Topographic Mapping of Active Faults in
Southern California with Satellite Optical Imagery, William
D Barnhart, Michael J Willis, Terryl L Bandy, Rich Briggs, Brianna
Morales, and Mark Faney (Poster Presentation 133)
Digital elevation models (DEMs) provide a core data set for many
studies of active tectonics. They allow for the identification and
quantification of fault offsets, modeling of localized and regional
fluvial responses to vertical and lateral fault displacements, and
mapping of active fault traces, among other applications. Very
high spatial resolution (<5 m per pixel)DEMs greatly enhance
these efforts by resolving subtle yet important variations in
topography. Moreover, DEMs and LiDAR observations with
WELCOME
2016 SCEC Annual Meeting page 155
spatial resolutions <5 m are increasingly used as geodetic tools
to measure co-seismic deformation. However, these high-
resolution data sets are commonly expensive to acquire,
encompass limited ground footprints, and are difficult to re-
acquire for the purposes of temporal change detection. High-
resolution satellite optical imagery catalogs provide one means
to overcome these limitations. Here, we present preliminary high
resolution DEMs (0.3-2 m per pixel) generated from in-track
stereo imagery of the WorldView satellite constellation. Our initial
work spans the Ventura and San Gorgonio Special Fault Study
Areas, as well as distributed faulting south of the San Jacinto and
Elsinore fault zones on trend with the 2010 M 7.2 El Mayor-
Cucapah rupture. DEMs were processed using the NASA Ames
Stereo Pipeline software suite and Ohio State University’s
Surface Extraction from TIN-Based Searchspace Minimization
(SETSM) algorithm on the University of Iowa’s Neon and
University of North Carolina’s KillDevil supercomputers. We also
present initial quality control metrics by comparing our derived
DEMs to existing topographic data sets, such as the B4 LiDAR
data set. Initial comparisons of our raw DEMs to B4 LiDAR
swaths indicate 1-sigma elevation uncertainties of ~1.5 m, and
we expect these uncertainties to decrease after co-registering
these data sets and applying various filtering algorithms. These
data sets and open-source software tools provide a means to
rapidly generate high-resolution DEMs of exceptional quality and
spatial resolution over broad regions, allowing for frequent re-
acquisition of data sets, geodetic analysis, and detailed
landscape analysis.
Inter-Period Correlations of SCEC Broadband Platform
Fourier Amplitudes and Response Spectra, Jeff R Bayless,
Paul G Somerville, Andreas Skarlatoudis, Fabio Silva, and
Norman A Abrahamson (Poster Presentation 292)
A principal intent of the SCEC BBP is for simulated time series to
be applied directly in engineering applications. To do so, the
inter-period correlations of epsilon need to be appropriate, since
these correlations are related to the width of peaks and troughs
in the spectra, and will therefore have an effect on the inelastic
response of structures. Epsilon is the normalized difference
between the observed (or predicted) Fourier Amplitude Spectra
(FAS) or Pseudo-Spectral Acceleration (PSA) and the mean
predicted natural log FAS or PSA.
We implement a validation scheme for the inter-period correlation
of epsilon for PSA and FAS of simulated ground motions on the
SCEC BBP. We calculate the correlations of epsilon for both FAS
and PSA, for each of the seven western United States Dreger et
al. (2013) validation events, for all available simulation methods
on the SCEC BBP, and compare with the observed correlation of
FAS from datasets such as NGA-West2 and NGA-East, and
compare with existing empirical models for PSA. To do this
calculation, we implement and verify computer codes on the
SCEC BBP for the calculation of the FAS, the smoothing of the
FAS, and the correlation of epsilon for FAS and PSA, including
simple FAS models from which epsilon values are derived.
PH5 for integrating and archiving different data types,
Bruce C Beaudoin, Derick Hess, and Steve Azevedo (Poster
Presentation 346)
PH5 is IRIS PASSCAL's file organization of HDF5 used for
seismic data. The extensibility and portability of HDF5 allows the
PH5 format to evolve and operate on a variety of platforms and
interfaces. To make PH5 even more flexible, the seismic
metadata is separated from the time series data in order to
achieve gains in performance as well as ease of use and to
simplify user interaction. This separation affords easy updates to
metadata after the data are archived without having to access
waveform data. To date, PH5 is currently used for integrating and
archiving active source, passive source, and onshore-offshore
seismic data sets with the IRIS Data Management Center (DMC).
Active development to make PH5 fully compatible with FDSN
web services and deliver StationXML is near completion. We are
also exploring the feasibility of utilizing QuakeML for active
seismic source representation.
The PH5 software suite, PIC KITCHEN, comprises in-field tools
that include data ingestion (e.g. RefTek format, SEG-Y, mseed,
and SEG-D), meta-data management tools including QC, and a
waveform review tool. These tools enable building archive ready
data in-field during active source experiments greatly decreasing
the time to produce research ready data sets. Once archived, our
online request page generates a unique web form and pre-
populates much of it based on the metadata provided to it from
the PH5 file. The data requester then can intuitively select the
extraction parameters as well as data subsets they wish to
receive (current output formats include SEG-Y, SAC, mseed).
The web interface then passes this on to the PH5 processing
tools to generate the requested seismic data, and e-mail the
requester a link to the data set automatically as soon as the data
are ready.
PH5 file organization was originally designed to hold seismic time
series data and meta-data from controlled source experiments
using RefTek data loggers. The flexibility of HDF5 has enabled
us to extend the use of PH5 in several areas one of which is using
PH5 to handle very large data sets. PH5 is also good at
integrating data from various types of seismic experiments such
as OBS, onshore-offshore, controlled source, and passive
recording. HDF5 is capable of holding practically any type of
digital data so integrating GPS data with seismic data is possible.
Since PH5 is a common format and data contained in HDF5 is
accessible randomly it has been easy to extend to include new
input and output data formats as community needs arise.
Earthquake source complexity? Near-fault velocity
spectra from laboratory failures and their relation to
natural ground motion, N. M Beeler, Brian D Kilgore, and
David A Lockner (Poster Presentation 029)
Natural earthquake sources have characteristic spectra, for
example at high frequency the displacement amplitude spectrum
of a body wave depends on frequency as 1 / f^2 [e.g., Hanks and
Wyss, 1972] and velocity amplitude as 1/f. Such natural
observations made at the Earth's surface are far removed from
the source and it is unknown whether these characteristic spectra
are due to the source or in part due to alteration along the source
to station path. Well-instrumented laboratory experiments may
provide some guidance in understanding the source
contributions to the frequency content of ground motion,
especially in instances where the surface and rheological
properties of the fault and the physical processes in the source
are well characterized and fault slip velocity is measured directly.
Here we report preliminary results on the velocity spectra of
WELCOME
2016 SCEC Annual Meeting page 156
laboratory failures: stick-slip on flat fault surfaces unconfined
(normal stress of 2 MPa), at intermediate confining pressure (100
MPa), at high confinement where shear melting occurs (300
MPa), slip on fractured surfaces at high confinement (300 MPa),
and rock fracture at intermediate confining pressure (50 MPa).
In contrast to a real earthquake, the theoretical on-fault velocity
amplitude spectrum of a laboratory failure event with constant
sliding friction follows 1/f^2. We compare our experiments to this
reference. The reference represents a laboratory faulting
experiment in which the fault properties make no contribution to
the observed spectrum at high frequencies. That is, the reference
is a fault that has no high frequency component: its strength does
not change with time during the event, thus does not to not
contribute to increasing or decreasing the slip speed, and which
contains no intrinsic source of high frequency energy, e.g., no
wear of the surfaces, no comminution of the existing shear zone
which can produce fracturing or other sources of small scale
rapid displacement that give rise to elastodynamic waves.
Among the notable experimental results are that flat fault
surfaces and intact rock failures are enriched in high frequency
motion relative to the reference and show the same characteristic
1/f decay as the velocity amplitude of natural earthquakes.
However, slip on rough fracture surfaces are depleted in high
frequencies relative to natural earthquakes, and are very similar
to the reference, essentially following the 1 / f^2 decay. Shear
melting shows evidence of mixed behavior, a 1/f decay at
intermediate frequencies and 1/f^2 at greater than 100 khz as if
the presence of melt inhibits very rapid changes in slip rate.
2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): Loss Analysis of
Earthquake Sequences, Sophia Belvoir, Vianca E Severino
Rivas, Jenepher M Zamora, Jadson da Silva Lima, Luis J
Gomez, Jozi K Pearson, Mark L Benthien, and Thomas H Jordan
(Poster Presentation 333)
The 2016 Undergraduate Studies in Earthquake Information
Technology (USE-IT) intern research program challenged the
Hazard and Risk Visualization Team to illustrate threatening and
probable multi-event earthquake scenarios in California using
ShakeMaps and risk analysis maps. The interns are
implementing FEMA’s Hazards United States (HAZUS) software
to estimate and visualize the potential losses including physical
damage of infrastructure, economic loss, and social impacts of
these multi-earthquake sequences. This is the first time multiple
events will be analyzed with HAZUS. As a team based project,
USEIT interns depend on collaboration with other working groups
within the intern program to identify potential earthquake
sequences. The Rate-State earthQuake Simulator (RSQSim)
was run on a high performance computer (HPC) to generate
catalogs of events. Rupture sequences were then selected to be
visualized in SCEC-VDO. Earthquake sequences that were seen
as most threatening had ShakeMap files created utilizing seismic
hazard analysis (OpenSHA) software. Multiple events were
combined as single ShakeMaps and imported into HAZUS for
analysis. Ground motion values were combined and calculated
in two different ways for multi-event ShakeMaps: 1) by using the
greatest value and 2) sum of the squares, where the maps
overlap. The greatest value method proves more appropriate for
events that are further apart geographically. The sum of the
squares method is ideal for events that are closer together. This
year interns can illustrate the most threatening multi-event
scenarios with hazard and risk maps providing colleagues and
the public with alternate scenarios to help create and revise
safety plans, guidelines, and economic strategies. In California,
most of the population is expecting the "Big One", but we are also
interested in scenarios that involve more than one large (M7+)
event within the same year, same month, same week, and even
the same day on the San Andreas Fault, as well as other
threatening southern California faults.
Internal structure of the San Jacinto fault zone at
Blackburn Canyon from a dense linear deployment across
the fault, Yehuda Ben-Zion, Pieter-Ewald Share, Amir A Allam,
Fan-Chi Lin, and Frank L Vernon (Poster Presentation 185)
We image the internal structure of the Clark section of the San
Jacinto fault zone (SJFZ) at Blackburn Canyon using teleseismic
and local earthquake waveforms recorded during about 1 month
by a linear array consisting of 125 three-component 5 Hz
geophones with an aperture of 2.4 km. The instrument spacing
is 10 m in a zone 400 m wide centered on the surface trace of
the fault, and ~30 m to the NE and SW of that zone. Analysis of
4 teleseismic events indicates an abrupt change in the horizontal
first motions of P phases around the central station, BS55,
coinciding with the surface trace. Variations in P arrival times
suggest a zone of slowness from station BS55 to ~350 m to the
NE, with maximum slowness ~130 m from BS55. Automatic
algorithms are used to detect P fault zone head waves (FZHW)
and S fault zone trapped waves (FZTW) generated by local
events. Over 200 FZHW candidate phases are detected for 57
M>1 events located SE of the array along the SJFZ. Inspection
of the picks reveals probable FZHW in most stations within ~1
km wide zone to the SW of station BS32 (~350 m NE from the
central station BS55). FZTW are detected for 36 M>1 events
located in a very broad region around the array. When
considering only stations in the central 400 m zone, 76% of the
FZTW detections are made for stations in the 200 m wide zone
NE of BS55. A 100 m wide internal region with 10 stations (B40
to B50), centered 100 m NE of BS55, has the largest amplitude
and lowest frequency for waveforms with FZTW, and most likely
overlies the core active damage zone of the Clark fault. Based
on these initial findings and geological constraints, the main Clark
fault at depth is likely closest to station BS55, damage is
asymmetric to the NE with most damage confined to a ~100 m
zone, and the damage zone terminates at a bimaterial interface
close to station BS32. Further analysis is needed to confirm the
discussed results and conclusions.
Sedimentation in Nearshore Basins as Related to
Paleoseismicity, Will Berelson, Alex Sessions, Josh West, and
James F Dolan (Poster Presentation 311)
The exceptional preservation and accumulation of Santa Barbara
Basin (SBB) laminae has been exploited by a trove of scientists
(Huselmann and Emery; Kolpack and Gorsline; Soutar;
Fleischer; Thornton; Drake; Schimmelmann et al.; Kennett et al.;
Behl et al. and many others) interested in resolving, at high-
temporal resolution, the sedimentation history of this coastal site.
Among the many curious aspects of SBB sedimentation is the
relative importance of hemi-pelagic vs. mass movement
contribution to the overall sediment package.
WELCOME
2016 SCEC Annual Meeting page 157
Visual characteristics of sediment has always been an important
observation made by sedimentologists. The olive-dark gray
package of SBB sediments, which upon closer inspection can
show mm or sub-mm scale coupling of lighter and darker
laminae, dominate Holocene sediments recovered in box, multi,
gravity and piston cores (Behl and Kennett ). Sedimentologists
who have worked on SBB cores have also noted the abundance
of gray layers, distinctly lighter in color and massive, compared
to the typical dark-olive laminated sediments. Described by
Huselmann and Emery ( ) as distal turbidite deposits, Fleischer’s
( ) analysis of their grain size and mineralogy resulted in the
interpretation that they formed from a suspension of silt-clay
sized material discharged during a major flood event.
Subsequent studies provided additional evidence in support this
mode of formation (Thornton, Behl, Stein) and complementary
work has tried to make the connection between flood layers and
climate/precipitation (Schimmelmann, Li).
This poster takes another view of the relevance of flood layers in
a deposit containing mainly hemi-pelagic laminated sediment.
We are particularly interested in how the frequency of gray layers
in SBB relate to paleoseismic events. Our first order
interpretation is that gray layers are related to both seismicity and
precipitation.
Rayleigh-wave ellipticity obtained from noise cross-
correlations across southern California, Elizabeth Berg,
Fan-Chi Lin, and Amir A Allam (Poster Presentation 199)
We analyze Rayleigh wave ellipticity, or Rayleigh wave H/V
(horizontal to vertical) amplitude ratios, computed from multi-
component ambient noise cross-correlations using over 300
stations throughout Southern California in 2015. Because
Rayleigh wave H/V ratios are sensitive to shallow local earth
structure, this method enables simple and accurate resolution of
near-surface geological features. For every station pair, we
simultaneously calculate the cross-correlation function in 1-day
time windows that are then stacked for all of 2015. Pre-
processing includes concurrent three-component temporal
normalization and spectral whitening in order to dependably
preserve amplitude information. Retaining correlation functions
with signal-to-noise ratios greater than five and two to three
wavelengths distance depending on period, we then measure
amplitude ratios between cross-correlations of different
components (radial-radial, radial-vertical, vertical-radial, vertical-
vertical) to determine Rayleigh wave H/V ratios at the two station
locations for each source-receiver noise cross-correlation pair.
The ratio between the radial-radial and vertical-radial
components (or radial-vertical and vertical-vertical) corresponds
to the H/V ratio of the source station while the ratio between
radial-radial and radial-vertical (or vertical-radial and vertical-
vertical) corresponds to the receiver station. For a given station,
we then compute the mean and the standard deviation of the
mean for all such measurements using every viable station pair,
yielding a single representative H/V ratio and its uncertainty for
that station. Our final maps of H/V ratios show relatively good
correspondence with known near-surface structure with a few
non-corresponding locations, motivating further study. High
ratios are observed in the LA Basin, Santa Barbara Basin, and
Salton Trough, while low ratios are noted in the San Gabriel, San
Jacinto and southern Sierra Nevada mountains. Unusually high
ratios seen in the southern Coast Ranges and Eastern California
Shear Zone may represent extremely local shallow sedimentary
features.
Sensitivity of deformation to activity along the Mill Creek
strand of the San Andreas fault within the San Gorgonio
Pass, Jennifer L Beyer, and Michele L Cooke (Poster
Presentation 060)
A significant portion of the deformation budget across irregular
fault networks can be accommodated by permanent off-fault
deformation that accumulates between earthquakes. This
deformation may correspond to a reduction in the slip rates along
the faults within these complex networks, which in turn may
contribute to uncertainty when evaluating seismic hazards within
southern California. We use well-validated 3D Boundary Element
Method (BEM) models to investigate the irregular fault geometry
through the San Gorgonio Pass. Within this region, on-going
debate centers on the activity level and geometry of the Mill
Creek strand of the San Andreas fault. Here we compare BEM
models with and without an active Mill Creek strand, while
investigating varying geometries of an active Mill Creek strand.
One alternate fault configuration consists of an active segment of
the Mission Creek fault transferring slip to the Galena Peak fault.
All models are run with a range of boundary conditions to account
for epistemic uncertainty of tectonic loading, providing a range of
slip rates and surface deformation. The presence of an active Mill
Creek strand decreases the dextral slip rates on other faults
within the San Gorgonio Pass; however, each model results in
fault slip rates that match several of the available geologic slip
rates. An active Mill Creek strand also produces much more
localized uplift, while the model without a Mill Creek strand
produces a more distributed uplift pattern in the San Bernardino
Mountains. For each fault geometry, we also compare two end
member models representing the interseismic loading of the
crust and deformation across multiple earthquake cycles. While
the interseismic models simulate the overall loading of faults
between earthquakes, the multi-cycle models simulate crustal
stresses that generate permanent off-fault deformation such as
evidenced by microseismicity. These end member models
produce distinct stressing rate patterns. The multi-cycle model
reveals that in regions of high fault complexity, the sum of the slip
on faults does not equal the plate velocity and a significant
portion of the strain is accommodated as distributed off-fault
deformation. Consequently, assessment of seismic hazards in
the southern California may be compromised due to insufficient
characterization of the complex fault geometry.
Performance-Based Fault Displacement Estimation With
Uniform California Earthquake Rupture Forecast 3
Probabilities, Glenn P Biasi (Poster Presentation 297)
We are implementing a performance-based engineering
approach for probabilistic fault displacement hazard analysis
(PFDHA) that builds on the third Uniform California Earthquake
Rupture Forecast (UCERF3). UCERF3 developed probabilities
of surface rupture for a state-wide system of active faults. The
USGS has adopted UCERF3 results for California, so its
implementation in PFDHA will be part of the next advance in
state-of-practice in California. This work specifically addresses
probability distributions for rupture displacement given fault
location.
WELCOME
2016 SCEC Annual Meeting page 158
Displacement estimates are developed in a staged approach that
preserves both best estimates and uncertainties. In the initial
stage, all ruptures crossing the site are gathered with their
magnitudes, lengths, rupture probabilities, and site location
within the rupture. Implied displacements given rupture can be
estimated using length vs. average displacement, and magnitude
vs. area scaling relations. Results are combined with UCERF3
rupture rates to form a displacement hazard curve. This estimate
is only a first approximation displacement hazard because
displacement variability within ruptures is not included.
To estimate rupture displacement variability we apply a database
of displacement measurements from historical ruptures. We find
that variability is not constant across ruptures, but depends on
where the fault crossing site is within the rupture, and may be
magnitude dependent. We are also evaluating whether variability
can be reduced using a statistical autoregressive approach.
These steps give a per-rupture displacement and uncertainty for
a crossing site. Two paths are considered to combine
displacement probabilities over multiple ruptures. The brute-
force method repeats per-rupture estimates for all UCERF3
ruptures that pass the site, and combines results using UCERF3
rupture probabilities. For some fault crossings such as the San
Andreas at the Los Angeles Aqueduct, this strategy would
involve tens of thousands of ruptures. In a second path, we first
extract the UCERF3 magnitude-frequency distribution (MFD)
from the site-crossing ruptures. Given the MFD, we vary the site
location within the ruptures. Three locations suffice for 6<M<7
ruptures, and of order 5 for larger events. We apply probabilities
in the site MFD to develop final displacement estimates and
uncertainties.
Seventy-two years of Surface Creep on the North
Anatolian fault at Ismetpasa: Implications for the southern
San Andreas and Hayward faults, Roger Bilham, Haluk
Ozener, David J Mencin, Asli Dogru, Semih Ergintav, Ziyadin
Cakir, Alkut Aytun, Bahadir Aktug, Onur Yilmaz, Wade Johnson,
and Glen S Mattioli (Poster Presentation 135)
Surface creep on the North Anatolian fault was first recognized
in 1969 in the form of an offset wall that had been constructed
across the fault 13 years after the 1944 Mw=7.4 Bolu/Gerede
earthquake. Publications by Ambraseys (1970) and Aytun
(1982), however, reported this offset as 24 cm and 18 cm
respectively resulting in an apparently unresolvable ambiguity in
1957-1944 fault creep rate. Sub-pixel measurements conducted
on high resolution photographs of the wall taken on the same day
in 1969 by Clarence Allen, however, reveal the true offset to have
been 12.5±0.5 cm, a value quite close to sub-pixel
measurements of offsets shown in photos published by the first
two authors. Ambraseys (1970) relates also that the nearby
railroad had been offset 1.5 m by the 1944 earthquake, and that
a subsequent repair was needed in 1950 to fix an additional
offset of 30 cm. The combined offset is almost a factor of three
smaller than the 5 m 1944 offsets of nearby trees, tracks, field
boundaries and aqueducts across the fault reported by Kondo et
al., (2005). These inconsistencies can be reconciled by
recognizing that Ambraseys' numbers refer to misalignment of
the railroad, and not to offset of the fault. When corrected for a
20° obliquity the 1944 coseismic offset increases to 3.7 m, and
pre-1950 afterslip increases to 0.74 m, that when summed with
50 cm of post 1969 creep, approximates the 1944 offsets
reported by Kondo et al. A carbon-rod creepmeter installed
across the fault in 2014 reveals that slip occurs episodically at a
mean decadal rate of ≈6 mm/yr (Figure 1), with creep confined to
the uppermost 4 km of the fault. Anecdotal reports indicate that
for some major North Anatolian Fault earthquakes coseismic slip
has been manifest as surface slip only after several hours. We
consider this may also have occurred at Ismetpasa and that
surface slip during the seismic cycle here may be considered
entirely aseismic. Models of the ≈270 year earthquake cycle
suggest that a steady background creep of 4.5 mm/yr now
dominates the afterslip process (Figure 2), and that a further ≈1
m of surface slip will precede the next Mw≈7.4 earthquake.
Surface creep events currently load the top of the seismogenic
zone, each 1.5-4.5 mm event incrementing strain by 1-2 E-06 in
a few hours. We consider it probable that a creep event will
contribute to future nucleation of seismic rupture between 5 and
14 km depths. The processes on the North Anatolian fault parallel
those on the Hayward and southern San Andreas faults.
A new way to estimate shear tractions on active faults in
southern California, Peter Bird (Poster Presentation 009)
Two kinds of thin-shell finite-element (F-E) model of quasi-static
neotectonics are available for southern California: (1) kinematic
(or “inverse”) F-E models which fit geodetic and geologic data by
weighted least-squares but contain limited physics and no stress
magnitudes; and (2) dynamic (or “forward”) physics-based F-E
models that build on crustal thickness, heat-flow, lab rheologies,
fault traces and dips and the momentum equation to predict
quasi-static stresses, strain-rates and fault slip-rates (but often
fail to match kinematic data). Potentially, fault slip-rates from a
good kinematic model can be used as targets for tuning a good
dynamic model, in at least two ways: (A) adjusting the effective
friction (in a quasi-static approximation) of each fault element;
and/or (B) adjusting horizontal shear tractions exerted on the
base of the lithosphere by mantle convection. To date, I have
adjusted friction (A) to tune a Shells dynamic F-E model of
neotectonics in southern California to better match long-term
fault slip-rates from the NeoKinema deformation model that was
used in UCERF3 and NSHM2014. The Shells dynamic solution
is revised 100 times, and in each revision the effective friction of
each of the 1000 fault elements is adjusted up or down by a small
amount (no more than ±0.01) based on the sense of the current
slip-rate error for that fault. Results show that over 60% of the
fault elements (including most of the San Andreas and Garlock
faults) march straight down to the lower effective-friction limit of
0.01 and stay there. Overall RMS error in fault slip-rates falls from
5.3 to 1.6 mm/a. Fault rakes were not imposed, but have
excellent realism. The obvious physical interpretation is that over
60% of active fault area in southern California experiences near-
total stress-drop in large earthquakes due to dynamic weakening
mechanisms like thermal pressurization or thermal
decomposition or fault melting. Thus, the peak pre-seismic shear
traction on most of these faults is not much more than the coming
stress-drop. Yet, other areas on the fault network retain higher
effective friction, concentrate shear tractions, and serve as
nucleation sites. However, results are still tentative and
preliminary because (a) potential input from the SCEC
Community Rheology Model has not yet been used; (b) potential
input from the SCEC Community Thermal Model has not yet
WELCOME
2016 SCEC Annual Meeting page 159
been used; and (c) method B of adjusting mantle shear tractions
has not yet been attempted.
How does fault geometry control earthquake magnitude?
Quentin Bletery, Amanda Thomas, Leif Karlstrom, Alan W
Rempel, Anthony Sladen, and Louis De Barros (Poster
Presentation 070)
Recent large megathrust earthquakes, such as the Mw9.3
Sumatra-Andaman earthquake in 2004 and the Mw9.0 Tohoku-
Oki earthquake in 2011, astonished the scientific community. The
first event occurred in a relatively low-convergence-rate
subduction zone where events of its size were unexpected. The
second event involved 60 m of shallow slip in a region thought to
be aseismicaly creeping and hence incapable of hosting very
large magnitude earthquakes. These earthquakes highlight gaps
in our understanding of mega-earthquake rupture processes and
the factors controlling their global distribution. Here we show that
gradients in dip angle exert a primary control on mega-
earthquake occurrence. We calculate the curvature along the
major subduction zones of the world and show that past mega-
earthquakes occurred on flat (low-curvature) interfaces. A
simplified analytic model demonstrates that shear strength
heterogeneity increases with curvature. Stress loading on flat
megathrusts is more homogeneous and hence more likely to be
released simultaneously over large areas than on highly-curved
faults.
Assessing evidence for connectivity between the San
Diego Trough and San Pedro Basin fault systems,
offshore Southern California., Jayne M Bormann, Graham M
Kent, Neal W Driscoll, and Alistair J Harding (Poster Presentation
109)
The seismic hazard posed by offshore faults for coastal
communities in Southern California is poorly understood and may
be considerable, especially when these communities are located
near long faults that have the ability to produce large
earthquakes. The San Diego Trough fault (SDTF) and San Pedro
Basin fault (SPBF) systems are active northwest striking, right-
lateral faults in the Inner California Borderland that extend
offshore between San Diego and Los Angeles. Recent work
shows that the SDTF slip rate accounts for ~25% of the 6-8
mm/yr of deformation accommodated by the offshore fault
network, and seismic reflection data suggest that these two fault
zones may be one continuous structure.
Here, we use hull-mounted Knudsen sub-bottom profiles, high-
resolution multichannel seismic (MCS) reflection profiles, and
legacy USGS and industry MCS profiles in combination with
multibeam bathymetric data to characterize recent deformation
on the SDTF and SPBF zones and to evaluate the potential for
an end-to-end rupture that spans both fault systems. The SDTF
offsets young sediments at the seafloor for ~130 km between the
US/Mexico border and Avalon Knoll. The northern SPBF has
robust geomorphic expression and offsets the seafloor in the
Santa Monica Basin. The southern SPBF lies within a 25-km gap
between high-resolution MCS surveys. Although there does
appear to be a through-going fault at depth in industry MCS
profiles, the low vertical resolution of these data inhibits our ability
to confirm recent slip on the southern SPBF.
Empirical scaling relationships indicate that a 200-km-long
rupture of the SDTF and its southern extension, the Bahia
Soledad fault, could produce a M7.7 earthquake. If the SDTF and
the SPBF are linked, the length of the combined fault increases
to >270 km. This may allow ruptures initiating on the SDTF to
propagate within 25 km of the Los Angeles Basin. At present, the
paleoseismic histories of the faults are unknown. We identify
three locations where thin lenses of sediment mantle the SDTF,
providing the ideal sedimentary record to constrain the timing of
the most recent event on the fault. Characterizing the
paleoseismic histories of the fault systems is a key step toward
defining the extent and variability of past ruptures, which in turn,
will improve maximum magnitude estimates for the SDTF and
SPBF systems.
Validation of ground-motion simulations using precarious
rocks, Elliot K Bowie, Mark W Stirling, and Chris Van Houtte
(Poster Presentation 290)
Validation of ground-motion simulations using precarious rocks
Elliot Bowie, Mark Stirling & Chris Van Houtte
We develop ground motion simulations for a major reverse fault
in central Otago, New Zealand, and use ancient precariously-
balanced rocks (PBRs; unstably balanced rocks on top of
pedestals) to validate the simulations, rather than the standard
approach of using instrumental strong motion records for
validation. The Dunstan Fault, a 60 km long reverse fault, is
responsible for the uplift of the Dunstan Mountains (1500-1600
m). PBRs are abundant within a few km of the southwestern end
of the fault, and are therefore conveniently located for validating
simulated ground motions from M>7 near-field Dunstan Fault
earthquakes. The fragility age (age since the PBR reached the
present unstable morphology), and fragility (the peak ground
acceleration (PGA) required to topple the PBR, based on field-
based estimates), are compared to the recurrence interval and
simulated ground motions of Dunstan Fault earthquakes. Earlier
studies show cosmogenic Be10 exposure dates for two PBRs are
in the range of 40,400 to 55,300 years B.P., and the Dunstan
Fault to show a recurrence interval of about 8000 years.
Therefore, the PBRs have likely experienced repeated large
earthquakes where ground-motions did not exceed their
fragilities (i.e. PGAs no greater than 0.7 g). The PBR fragilities
fall within the range of PGAs produced by the simulations (0.16-
1.33 g), with about 16% of them exceeding the highest fragility.
Decreasing kappa from the default value of 0.04 to a value more
representative of Otago (0.016) results in an overall increase of
simulated PGAs by about 30%, with 36% of them exceeding the
highest PBR fragility (0.7 g). We are currently identifying the
parameters responsible for the high PGAs in the 36% of
simulations. Our research represents the first effort at using
PBRs to validate ground-motion simulations in New Zealand, and
has been jointly supported by University of Otago, QuakeCoRE,
and GNS Science.
Multi-scale Structural Characterization of the Mecca Hills
Fault System in the NE block of the Southern San
Andreas Fault System, California, Kelly K Bradbury, Amy C
Moser, Sarah A Schulthies, and James P Evans (Poster
Presentation 110)
WELCOME
2016 SCEC Annual Meeting page 160
We examine the structural architecture, mineralogy, and
alteration of fault zones across a range of scales in the Mecca
Hills in southern California. Our research group is focusing on
outcrop-scale field and laboratory measurements of 4 well-
exposed NNE trending subsidiary SSAF faults (Painted Canyon,
Platform, Eagle Canyon, and the Hidden Springs Fault) that
exhibit a range of lengths and displacements, and that formed in
the same tectonic regime, at nearly the same time, east of the
main SSAF.
Faults exhibit a bimodal distribution of strike-slip and dip-slip slip
vectors. NE-SW striking strike-slip faults are at high angles to the
main traces of the major dextral faults, whereas dip-slip faults
strike E-W. These data imply separate structures accommodate
strike-slip and dip-slip motion and transfer slip.
We determine the nature and distribution of deformation features
within the damage zones of each fault, the micro- to nano-scale
composition and texture of the protolith and fault-related rocks,
and the alteration signatures present within each fault zone.
Complex fault geometries with multiple, anastomosing strands
that dip <90° and asymmetric damage zones are observed.
Several m’s thick gouge zones and 10’s of meters wide damage
zones occur. Numerous highly reflective silica-rich slip surfaces
are identified and exhibit textures suggestive of vapor phase-
liquid partitioning during slip. Matrix-supported brecciation and
cataclastic flow textures imply mineralization associated with
rapid fault slip at these localities. Iron oxide mineralization is also
associated with cataclasis at several localities.
Two predominate phases of veins are consistently detected
within fault damage zones and suggest hydrothermal alteration
in the Mecca Hills region: 1) a N-trending quartz ± calcite ± zeolite
system, typically oriented parallel to faulting &/or foliation; and 2)
an EW-trending system composed of quartz ± calcite ±
laumontite ± palygorskite ± trace sulfides & metals. Quartz vein
deformation microstructures display evidence for multiple stages
of brittle deformation.
Petrographic, SEM, and geochemical results imply that fault
rocks in the Mecca Hills damage zones did not form from simple
mechanical grinding of the protolith. For example, a relative
decrease in SiO2, increase in CaO, and increase in FeO occurs
between the protolith and the fault-related rocks. Barium
concentrations also decrease significantly from the protolith to
the fault-related rocks, suggesting hydrothermal fluids
preferentially stripped Ba from the system.
These data, coupled with our pre-existing work on the San
Andreas fault address: 1) the differences in the physical
properties between a major plate boundary fault and smaller
scale subsidiary faults; and 2) provides critical information on the
physical and chemical processes active during fault zone
deformation in a region of triggered slip and creep.
Guidance on the utilization of ground motion simulations
in engineering practice, Brendon A Bradley, Didier Pettinga,
and Jack W Baker (Poster Presentation 296)
This poster provides an update on current progress to develop
guidance on the utilization of ground motion simulations in
engineering practice. A summary is provided of the key
ingredients of ground motion simulations and the critical role of
verification, validation and utilization documentation in order to
develop confidence in the predictive capabilities of the
simulations. The conceptual framework in which validation can
be undertaken is examined in particular. An example validation
process for simulations of a prospective Alpine Fault rupture is
illustrated for clarity.
Recent Advances of the ADER-DG Finite Element
Method for Seismic Simulations, Alexander N Breuer,
Alexander Heinecke, and Yifeng Cui (Poster Presentation 342)
In this poster we present a novel local time stepping (LTS)
scheme for ADER-DG finite element method and a holistic
optimization targeting seismic simulations on the Intel Xeon Phi
x200 processor, codenamed Knights Landing (KNL).
The ADER-DG method’s flexibility has proven to be highly
valuable when simulating multi-physics earthquake simulations
coupling seismic wave propagation and the rupture process.
While LTS offers great opportunity to reduce the computational
footprint of ADER-DG, complex update-dependencies of the
elements might occur if implemented naively. Our LTS-scheme
summarizes elements with similar time steps and thus aims for
optimal large-scale throughput by trading some of the algorithmic
optimality.
Our results include petascale performance when using local time
stepping for the computationally heavy seismic wave
propagation. Further, our performance comparisons
demonstrate that KNL is able to outperform its previous
generation, the Intel Xeon Phi coprocessor x100 family, by more
than 2.6× in time-to-solution when running LTS-workloads.
Additionally, our results show a 2.9× speedup compared to latest
Intel Xeon E5v3 CPUs.
We conclude our poster by presenting recent work, aiming at
even higher machine utilization by fusing multiple earthquake
simulations into one single execution of the ADER-DG solver.
This approach greatly reduces the memory requirements over
sequential execution, since a significant fraction of the data is
shared among the fused simulations. In addition, our approach
does not require zero-operations to fulfill hardware requirements
in the micro-kernels. Our performance benchmarks (4th order,
matrix Fm2, M=9, N=20, K=20, 16% non-zero, hot L2) underline
the potential of the approach and show that we are able reduce
time-to-solution by more than 3.5x on the Intel Xeon Phi
processor if compared to the fastest non-fused implementation.
Higher Earthquake Intensity Attenuation Rates in the
Urbanized Southern Puget Lowland Than Elsewhere
Along the Cascadia Forearc, Thomas M Brocher (Poster
Presentation 280)
The attenuation of seismic intensity with distance for 15
magnitude M4.8 to M6.8 earthquakes between 1949 and 2015
shows significant variation along the Cascadia forearc. Felt
intensities for these earthquakes were taken from NOAA’s U.S.
Earthquake Intensity Database (1635-1985), USGS Open-File
Reports, and a 2015 Did You Feel It? report. I follow the
approach of Bakun et al. (2002), and fit linear intensity-
attenuation functions to the median distances at which the
earthquakes were reported felt. With the exception of the
coefficients of the distance-dependent term, I used the same
form-of-equation and constants as used by those authors.
WELCOME
2016 SCEC Annual Meeting page 161
Attenuation rates for 4 M5.1 to M5.8 crustal earthquakes in the
Cascadia forearc south of Puget Lowland are well-resolved and
lie between 0.018 and 0.023 ± 0.002 MMI/km, close to that
determined by Bakun et al. (2002) and Bakun and Wentworth
(1997) for crustal earthquakes west of the Cascades and in
California. In contrast, poorly resolved attenuation rates for 4
M4.8 to M5.2 crustal earthquakes within Puget Lowland lie
between 0.032 and 0.038 ± 0.005 MMI/km, 45 to 72% larger.
Attenuation rates for 5 M4.8 to M6.8 inslab earthquakes beneath
Puget Lowland and the southeastern end of Vancouver Island
are well-resolved and show no dependence on magnitude: they
range between 0.0205 and 0.030 ± 0.002 MMI/km, intermediate
between these two crustal rates. The attenuation rate for one
inslab event, the M5.8 1999 Satsop earthquake located to the
west of the Lowland, was noticeably higher at 0.034 ± 0.003
MMI/km. Although magnitude dependence of the attenuation
rates may explain some of the observed variations, well-resolved
higher attenuation rates of felt intensities for 3 M5.4 to M6.8
crustal and inslab earthquakes in southern Puget Lowland are
associated with the presence of large glaciated Cenozoic basins
and with the northward thinning of the Crescent Formation, a unit
having high seismic velocities and presumably high Q.
The Action of water films at Å-Scales in the Earth:
Implications for Midcrustal Eathquakes and
Overpressuring, Kevin M Brown, Dean Poeppe, and IODP Leg
348 Shipboard Party (Poster Presentation 037)
Water properties change with confinement within nanofilms
trapped between natural charged clay particles. We investigated
nanofilm characteristics utilizing sediments from deep drilling of
the Nankai subduction zone at Site C0002 of the Integrated
Ocean Drilling Program (IODP) through high-stress laboratory
compression tests in combination with analyses of expelled pore
fluids. We show that below 1-2 km, there should be widespread
ultrafiltration of migrating fluids. Experiments to >~100 MPa
normal compression collapses pores below a few ion monofilm
thicknesses. A reduction towards a single condensing ion
monofilm <10-20 Å occurs as stresses rise >100-200 MPa. Thus,
porosity in high mineral surface area systems only consists of
double and single monofilms at depths below a few km. It is not
clear that thermal pressurization mechanisms and associate
dynamic weakening mechanism operate effectively in monofilms.
The resulting semipermeable properties also result in active flow
causing variable segregation of ions and charged isotope
species, increasing salinities in residual pore fluids at depth, and
freshening of the fluids expelled during consolidation.
Retardation of nanofilm collapse at ~17Å or ~2 x monolayer
thickness below 2km at Nankai, however, indicates the onset of
substantial geopressuring on the deeper “tremor” generating
seismogenic fault. The properties of strange monofilm water,
thus, have considerable implications for the deep hydrology of
major Mw 8+ earthquake generating subduction zones. Indeed,
the combined effects of advective flow, ultrafiltration, and
diffusion and diagenesis could provide a unifying explanation for
the origins of overpressuring and geochemical signals observed
in many natural systems.
Using deformation rates in Northern Cascadia to
constrain time-dependent stress- and slip-rate on the
megathrust, Lucile Bruhat, and Paul Segall (Poster
Presentation 161)
In northern Cascadia, estimates of interseismic slip deficit on the
megathrust are complicated by the poorly understood “gap”
between the down-dip limit of the locked region and the top of the
Episodic Tremor and Slip (ETS) zone, as revealed by kinematic
inversions of deformation rates. Physics-based models of slow
slip events (SSE) match the average ETS displacements, and
predict nearly constant shear stress when averaged over many
SSEs. Yet, when the fault is locked to the top of the ETS zone,
such models predict too much locking when compared to
observed decadal-scale velocities, especially uplift rates.
Minimum norm inversions of GPS and tide gauge/leveling rates
show that larger slip rates are necessary within both the gap and
the ETS zone relative to the model with constant shear stress,
requiring non-stationary shear stress rates. We invert for the
distribution of shear stress rate on the fault below a locked zone
that best fits the data. We find that negative shear stress rates
within both the gap and the ETS region, reaching -3kPa/yr at a
depth of 25km, and locking depths around 21km are required by
the data.
To explain the negative shear stress rates, we investigate models
where both the size of the creeping region and the slip profile
evolve with time. We explore whether the gap and ETS region
act a quasi-static crack, potentially driven up-dip by plate motion
at the down dip limit of the ETS region. Stress reaches a peak at
the crack tip, and then decreases over a cohesive region. Using
MCMC we solve for the locking depth, the crack friction
parameters, and the up dip propagation velocity. Best fitting
models reveal that a very slow up dip propagation is needed to
fit the data and that the gap seems to act as a region of fault
weakening. Possible explanations for this include a slow
decrease in normal stress within the gap, possibly due to an
increase in pore pressure, or a long-term reduction in fault
friction.
Toppling of PBRs Exposed to Ground Motions Estimated
from the Composite Source Model of Earthquakes,
Richard J Brune, John G Anderson, Glenn P Biasi, and James N
Brune (Poster Presentation 291)
We have improved the description and documentation of
precariously balanced rocks (PBRs) at Lovejoy Buttes 15 km NE
of the San Andreas fault at the western edge of the Mojave
Desert, and at sites within 20 km of the San Jacinto fault near the
trans-tensional stepover between the Claremont and Casa Loma
sections. We compare PBRs fragilities to ground motion
simulations based on the composite kinematic source model of
Anderson (SRL 2015), which includes effects of earth structure,
attenuation, scattering, sub-event stress drops, and random
distributions of slips. Along the San Jacinto we used three rupture
models: a single rupture along the length of the two segments
combined, and separate ruptures along each of the Claremont
and Casa Loma segments. For each rupture we used models
with both 20 and 60 bar sub-event stress drop.
Source models with the 20 bar sub-event stress drop are mostly
consistent with the PBRs at Lovejoy and San Jacinto not having
been knocked down, while several rocks at both locations are
likely to be toppled by the 60 bar synthetics, thus potentially
WELCOME
2016 SCEC Annual Meeting page 162
constraining the modeling parameter. If we were to reject 60
bars, then the 20 bar model would tell us that we cannot reject a
continuous rupture through the stepover on the basis of the
PBRs. For a given PBR, the probability of toppling appears to be
sensitive to the distance to the nearest large sub-event in a
simulation. The 20 bar synthetics tend to be comparable to or on
the low side of GMPEs depending on the oscillator period.
Whether or not a particular PBR is toppled depends on using the
actual 3d directionality (polarization) of the ground motion
simulations as well as the correct shape and orientation. The
ground motions that do not topple modeled rocks in their actual
orientation are found to more often topple the same rocks if their
simulated orientations are rotated by 90 degrees. Thus the actual
orientation of the PBRs and the use of synthetics incorporating
the physics of source radiation patterns and wave propagation is
essential for somewhat finely-tuned limits on pre-historic near-
source ground motions.
Testing for slip rate changes on the Sierra Madre fault:
Progress on dating an offset terrace surface of possible
middle Pleistocene age, Reed J Burgette, Nathaniel Lifton,
Katherine M Scharer, Devin McPhillips, and Austin Hanson
(Poster Presentation 099)
The Sierra Madre fault (SMF) system juxtaposes the San Gabriel
Mountains against a series of basins along the northern margin
of the Los Angeles metropolitan area. Previous studies have
suggested that deformation rates on this fault vary spatially as
part of the broader plate boundary system, or that the locus of
deformation has migrated southward into the Los Angeles basin
over the Quaternary. A well-preserved flight of fan terraces in the
Arroyo Seco area of Altadena and Pasadena, CA provides a
location to assess whether there is a long-term reduction in strain
on the SMF. A complementary slip-rate study (Hanson et al., this
meeting) focused on lower fan terraces that are offset vertically
~3-20 m, providing late Pleistocene slip rates for this fault. This
project targets the prominent Gould Mesa geomorphic surface
thought to be middle Pleistocene in age, locally preserved across
the Central SMF. The Gould Mesa surface is capped by a thick,
deep red soil, and is thought to be ~200-500 ka based on
published regional soil correlations. Nearby wells in the footwall
of the SMF contain a thick red clay, interpreted to be correlative
with this surface soil, and constrain the vertical separation of the
top of the Gould Mesa surface across the frontal strands of the
fault zone to be ~250 m. We are applying cosmogenic nuclide
isochron burial dating to determine the age of the upper part of
the Gould Mesa alluvial deposit. Burial dating analyzes pairs of
radionuclides (10Be and 26Al in this case) produced by high-
energy cosmic rays in clasts at the surface (and potentially during
transport) that have subsequently been deeply buried. Decay of
the radionuclides at different rates makes their ratio sensitive to
duration of burial, and analysis with an isochron technique yields
ages since burial. We collected quartz-rich cobbles from natural
and roadcut exposures from 8-10 m below the upper surface of
Gould Mesa at two locations. Each set of samples was collected
from a layer <50 cm thick to ensure the clasts have had a
common burial history. Physical sample processing is complete,
and the samples are now undergoing chemical processing and
analysis. Based on synthetic modeling of the burial depth and
inheritance values from other cosmogenic studies on the SMF,
we anticipate age uncertainties for each of the two isochron burial
dating sites to be ~70 ka or better, which we hope will reduce the
uncertainties on the long-term Quaternary slip rate at this site by
a factor of ≥2 compared to the current level used by UCERF3.
2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): Documentary and
Virtual Reality Game, Geneva I Burkhardt, Aadit R Doshi,
Emily R Olmos, Emma L Huibregtse, Drew P Welch, Mark A
Romano, Jason Ballmann, Jozi K Pearson, and Mark L Benthien
(Poster Presentation 334)
As a liaison between scientists and the public, the Media Team
translates complex concepts into accessible information. Its
documentary explains concepts which fellow interns have
researched and developed, illustrates the program’s impact on
interns, describes the collaboration of interns amongst
themselves and with mentors, and conveys how USEIT research
applies to safety. The 2016 Grand Challenge, which involves
earthquake forecasting, will provide a better understanding of risk
and hazard probability in seismic events. The interns’ answer to
the Grand Challenge has the potential to alert the public to the
probability and danger of earthquakes, but only if the research is
well communicated. The documentary will also be used to
encourage funding. This is essential to the continuation of the
USEIT program and the extension of earthquake research. To
emphasize USEIT’s value, the Media Team conducted
interviews. Here, interns explained the program’s impact on their
careers, learning paths, and lives. In addition to a documentary,
the Media Team focused on applying interactive media to
earthquake science. The game design aspect of the team
teaches earthquake safety through immersive technology. A
major goal of earthquake science is to inform the public on what
they should do during an earthquake. In some cases, the
suddenness of the event causes even the most experienced
people to forget the correct procedure. A Virtual Reality game
offers a fun, risk-free environment for the public to train and learn
what to do during an earthquake. The game balances realism
and game design liberty to create an optimal environment for
players to maximize the learning experience. Players will be
concerned about glassware, debris, and furniture that could
potentially hurt them. The game trains the player in three
common scenarios: a home, a school, and an office. The player
will then play the game to practice thinking quickly in unfamiliar
scenarios and to find a safe location to take cover for the duration
of the earthquake. Each level in Play mode is randomly
generated, providing the player with infinite scenarios to practice.
The game group also experimented with Augmented Reality to
illustrate faults and earthquake events through an app for phones
or tablets. The Media Team is essential to USEIT because
without communication, science is difficult to apply to other fields,
nor can the importance of USEIT be understood by the public.
Using Simulated Ground Motions to Constrain Near-
Source Ground Motion Prediction Equations in Areas
Experiencing Induced Seismicity, Samuel A Bydlon, and Eric
M Dunham (Poster Presentation 268)
Recent increases in seismic activity in historically quiescent
areas such as Oklahoma, Texas, and Arkansas, including large,
potentially induced events such as the 2011 Mw 5.6 Prague, OK,
earthquake, have spurred the need for investigation into
expected ground motions associated with these seismic sources.
WELCOME
2016 SCEC Annual Meeting page 163
The neoteric nature of this seismicity increase corresponds to a
scarcity of ground motion recordings within ~50 km of
earthquakes Mw 3.0 and greater, with increasing scarcity at
larger magnitudes. Gathering additional near-source ground
motion data will help better constraints on regional ground motion
prediction equations (GMPEs) and will happen over time, but this
leaves open the possibility of damaging earthquakes occurring
before potential ground shaking and seismic hazard in these
areas are properly understood. To aid the effort of constraining
near-source GMPEs associated with induced seismicity, we
integrate synthetic ground motion data from simulated
earthquakes into the process. Using the dynamic rupture and
seismic wave propagation code waveqlab3d, we perform
verification and validation exercises intended to establish
confidence in simulated ground motions for use in constraining
GMPEs. We verify the accuracy of our ground motion simulator
by performing the PEER/SCEC layer-over-halfspace comparison
problem LOH.1 Validation exercises to ensure that we are
synthesizing realistic ground motion data include comparisons to
recorded ground motions for specific earthquakes in target areas
of Oklahoma between Mw 3.0 and 4.0. Using a 3D velocity
structure that includes a 1D structure with additional small-scale
heterogeneity, the properties of which are based on well-log data
from Oklahoma, we perform ground motion simulations of small
(Mw 3.0 - 4.0) earthquakes using point moment tensor sources.
We use the resulting synthetic ground motion data to develop
GMPEs for small earthquakes in Oklahoma. Preliminary results
indicate that ground motions can be amplified if the source is
located in the shallow, sedimentary sequence compared to the
basement. Source depth could therefore be an important variable
to define explicitly in GMPEs instead of being incorporated into
traditional distance metrics. Future work will include the addition
of dynamic sources to develop GMPEs for large earthquakes.
CyberShake Advancements: Broadband Seismograms
and Central California Sites, Scott Callaghan, Philip J
Maechling, Christine A Goulet, Karan Vahi, Robert W Graves,
Kim B Olsen, Kevin R Milner, David Gill, Yifeng Cui, and Thomas
H Jordan (Poster Presentation 337)
The CyberShake computational platform, developed by SCEC,
is an integrated collection of scientific software and middleware
that performs 3D physics-based probabilistic seismic hazard
analysis (PSHA). CyberShake integrates large-scale parallel and
high-throughput research codes to produce probabilistic seismic
hazard curves for individual locations of interest and hazard
maps for an entire region.
The most recent CyberShake deterministic hazard models have
been calculated at seismic frequencies up to 1 Hz. To expand
this with PSHA results at frequencies of wider interest to civil
engineers (1-10 Hz), we performed CyberShake Study 15.12,
which augmented the 1 Hz deterministic results using high-
frequency stochastic seismograms produced using the Graves
and Pitarka (2015) module from the SCEC Broadband Platform.
For each of about 500,000 events and at each of 336 locations
of interest, stochastic seismograms were simulated, filtered,
combined with deterministic seismograms, and post-processed
to obtain intensity measures and hazard curves at frequencies
from 0-10 Hz. Additionally, Study 15.12 included the calculation
of 9 new duration measures for each simulated earthquake of
interest to building engineers.
In developing and verifying CyberShake, we have focused our
modeling in the greater Los Angeles region. We are now
expanding the hazard calculations into Central California. Using
workflow tools running jobs across two large-scale open-science
supercomputers, NCSA Blue Waters and OLCF Titan, we are
calculating 1 Hz deterministic PSHA results for over 400
locations in Central California. For each location, we are
producing hazard curves using two different velocity models: a
3D central California velocity model created via tomographic
inversion (CCA), and a regionally averaged 1D model. These
new results will provide low-frequency ground motion
exceedance probabilities for the rapidly expanding metropolitan
areas of Santa Barbara, Bakersfield, and San Luis Obispo. The
results will also lend new insights into the effects of directivity-
basin coupling associated with basins juxtaposed with major
faults such as the San Andreas, and produce physics-based
PSHA results for structures of interest such as water pumping
stations.
We will present our approach for incorporating stochastic
seismograms into CyberShake, show results from Study 15.12
as well as preliminary results for new Central California locations,
and describe our future CyberShake plans.
The most recent CyberShake deterministic hazard models have
been calculated at seismic frequencies up to 1 Hz. To expand
this with PSHA results at frequencies of wider interest to civil
engineers (1-10 Hz), we performed CyberShake Study 15.12,
which augmented the 1 Hz deterministic results using high-
frequency stochastic seismograms produced using the Graves
and Pitarka (2015) module from the SCEC Broadband Platform.
For each of about 500,000 events and at each of 336 locations
of interest, stochastic seismograms were simulated, filtered,
combined with deterministic seismograms, and post-processed
to obtain intensity measures and hazard curves at frequencies
from 0-10 Hz. Additionally, Study 15.12 included the calculation
of 9 new duration measures for each simulated earthquake of
interest to building engineers.
In developing and verifying CyberShake, we have focused our
modeling in the greater Los Angeles region. We are now
expanding the hazard calculations into Central California. Using
workflow tools running jobs across two large-scale open-science
supercomputers, NCSA Blue Waters and OLCF Titan, we are
calculating 1 Hz deterministic PSHA results for over 400
locations in Central California. For each location, we are
producing hazard curves using both a 3D central California
velocity model created via tomographic inversion (CCA), and a
regionally averaged 1D model. These new results will provide
low-frequency ground motion exceedance probabilities for the
rapidly expanding metropolitan areas of Santa Barbara,
Bakersfield, and San Luis Obispo. The results will also lend new
insights into the effects of directivity-basin coupling associated
with basins juxtaposed with major faults such as the San
Andreas, and produce physics-based PSHA results for
structures of interest such as water pumping stations.
We will present our approach for incorporating
stochastic seismograms into CyberShake, show results from
Study 15.12 as well as preliminary results for new Central
California locations, and describe our future CyberShake plans.
WELCOME
2016 SCEC Annual Meeting page 164
Dynamic friction in sheared fault gouge: implications of
acoustic vibration on triggering and slow slip, Jean M
Carlson, Charles K Lieou, and Ahmed E Elbanna (Poster
Presentation 039)
Friction and deformation in granular fault gouge are among
various dynamic interactions associated with seismic
phenomena that have important implications for slip mechanisms
on earthquake faults. To this end, we propose a mechanistic
model of granular fault gouge subject to acoustic vibrations and
shear deformation. The grain-scale dynamics is described by the
Shear-Transformation-Zone theory of granular flow, which
accounts for irreversible plastic deformation in terms of flow
defects whose density is governed by an effective temperature.
Our model accounts for stick-slip instabilities observed at seismic
slip rates. In addition, as the vibration intensity increases, we
observe an increase in the temporal advancement of large slip
events, followed by a plateau and gradual decrease.
Furthermore, slip becomes progressively slower upon increasing
the vibration intensity. The results shed important light on the
physical mechanisms of earthquake triggering and slow slip and
provide essential elements for the multiscale modeling of
earthquake ruptures. In particular, the results suggest that slow
slip may be triggered by tremors.
Slip and Seismic Radiation Along Bi-material Faults: An
Experimental Analysis, Brett M Carpenter, Ximeng Zu,
Xiaowei Chen, and Ze'ev Reches (Poster Presentation 027)
Large displacements along faults frequently juxtapose crustal
blocks of different compositions and mechanical properties. Such
juxtapositions have been termed bi-material faults (BFs). It has
been theoretically shown that a BF setting may strongly affect
rupture characteristics, energy radiation and damage during
earthquakes. In order to gain a better understanding of BFs, we
designed laboratory experiments on rock samples at conditions
similar to earthquake rupture, and present the preliminary results
of the mechanical and acoustic characteristics of experimental
bi-material faults.
Our experimental faults are composed of contrasting rock pairs
of a stiff, igneous block (diorite, granite, gabbro) sheared against
a more compliant, sedimentary block (sandstone, limestone,
dolomite). The experiments were conducted on a rotary
apparatus at velocities up to 0.1 m/s and a normal stress of to
0.5 MPa. The experimental faults were loaded under both
constant velocity, relevant to earthquake rupture propagation,
and under power-density control. Fault behavior was monitored
with stress, velocity and dilation sensors, as well as four 3D
accelerometers.
In our preliminary work, we observed differences in mechanical
behavior based on the bi-material composition of the fault. We
find greater fault parallel deformation and wear in experiments
between sedimentary rock types, when compared with those that
incorporated igneous rock types. Under power-density loading,
we observed differences in the initiation of sliding, stick-slip
versus stable sliding, in experiments conducted on a
sandstone/gabbro pair and a diorite/diorite pair respectively.
These differences were manifested in distinct disparities in the
evolution of velocity with slip, which would have important
implications for rupture propagation along these faults. We also
observed difference in seismic radiation based on the bi-material
composition of the fault. In general, we recorded higher
amplitude, more impulsive, and more frequent events through
igneous rocks, whereas in sedimentary rocks, we recorded lower
amplitude, more emergent, and fewer events, as well as tremor
like signals. Continuing work is aimed relating our mechanical
and acoustic observations to earthquake parameters, e.g.,
seismic radiation, rise-time, and stress drop, during slip along bi-
material faults.
Fault-Zone Exploration in Highly Urbanized Settings
Using Guided Waves: An Example From the Raymond
Fault, Los Angeles, California, Rufus D Catchings, Janis L
Hernandez, Robert R Sickler, Mark R Goldman, Joanne H Chan,
and Coyn J Criley (Poster Presentation 257)
Locating faults in the near-surface beneath highly urbanized
settings is challenging because urbanization obscures
geomorphic evidence of the faults in relatively short time periods.
Yet, urban faults can represent major hazards because they
directly underlie large populations. Such is the case along the
western Raymond Fault in Los Angeles, where it steps over to
the Hollywood Fault. It is important to identify the main and
auxiliary traces that may underlie homes and buildings.
Paleoseismic trenching and other invasive techniques offer the
clearest evidence of faulting, but long trenches cannot be
economically excavated in highly urbanized areas or zones
where faulting is widely distributed, such as fault step overs. In
May 2016, we used an alternative, non-invasive approach to
locate faults between the Raymond and Hollywood faults. This
technique involves using peak ground velocities (PGV)
measured from guided-wave energy generated within a known
fault trace, as recently done elsewhere (Catchings et al., 2013).
From a known site (from trenching) of the Raymond Fault, we
used an accelerated weight drop to input seismic energy into the
fault zone. The seismic energy was recorded by seismic arrays
located about 400 to 700 m from the source and aligned
approximately perpendicular to the expected trend of the
Raymond Fault. Although the area was bisected by a four-lane
highway and numerous busy streets, our recorded seismic
energy was largely noise free as a result of stacking more than
200 shots. We observed four distinct zones of high PGV values
along the array, none of which coincided with roadways or other
local noise sources. Instead, these zones of high PGV coincided
with fault traces inferred from geologic mapping, borings, and
pre-urbanization (1923; 1928) aerial photos. We find that the
guided-wave PGV method is highly effective in locating faults in
both urban and rural areas.
Earthquake cycles on rate-state faults: how does
recurrence interval and its variability depend on fault
length? Camilla Cattania, and Paul Segall (Poster Presentation
044)
The concept of earthquake cycles consisting of periodic events
on the same patch is often invoked or assumed when discussing
seismic risk. However, large faults exhibit more complex
behavior than periodic stick-slip cycles. Some events, such as
the 2004 Parkfield earthquake, are delayed relative to the
average recurrence interval; in other cases, the rupture area is
larger or smaller than expected. In contrast, small earthquakes
can be very predictable: locked patches surrounded by aseismic
WELCOME
2016 SCEC Annual Meeting page 165
creep can rupture periodically in events with nearly identical
waveforms.
Here we use numerical tools and ideas from fracture mechanics
to explore the factors determining recurrence interval (T), rupture
size and their variability at different scales. On rate-state faults,
T has been estimated by assuming a stress drop given by the
difference of steady-state stress at coseismic and interseismic
slip speeds, and stressing rate inversely proportional to fault
length. However, Werner and Rubin (2013) showed that the
scaling of T with fault size in numerical simulations is better
explained by an energy criterion: on faults loaded from below, full
ruptures occur when the elastic energy release rate at the top of
the fault reaches the fracture energy. We run simulations of
seismic cycles on rate-state faults, including dynamic weakening
due to thermal pressurization. We model a fault loaded from
below at constant slip velocity divided in a velocity-weakening
section over a velocity strengthening one. We find that T
increases with thermal pressurization, and we verify that the
energy argument, modified to account for the fracture energy
from thermal pressurization, provides a good estimate of T and
its scaling with fault size.
We suggest that the recurrence interval is determined by two
timescales: the time required to accumulate sufficient elastic
energy for full rupture (Tf), and the nucleation time, controlled by
the propagation of a creep front into the velocity-weakening
region (Tn). Both timescales depend on fault length: Tf increases
with length, while Tn decreases. The latter is due to the fact that
the velocity-strengthening region experiences faster afterslip on
larger faults.
When Tn < Tf, partial ruptures occur; for large enough faults, Tn
<< Tf multiple partial ruptures are observed. This causes more
heterogeneous stress fields and more variability in the
recurrence interval. Therefore we suggest that fault size alone,
by determining the difference in time scales for nucleation and
full rupture, can lead to more variability in seismic behavior
without requiring heterogeneity in frictional properties or
background stress.
Fault Scarp Degradation Analysis At Dragon’s Back Using
High Resolution Topography Data, Emil Chang, Gilles
Peltzer, and Seulgi Moon (Poster Presentation 079)
Fault scarps degrade over time, and their geomorphic forms may
provide information of relative activity of surface processes or
formation ages. The recent advances in remote sensing
techniques and high-resolution topographic data allow us to
quantify accurate geomorphic ages of fault scarps and to
examine their spatial variations within a fault zone. Our goal was
to quantify relative geomorphic ages, assess the quality of
different types of high-resolution data, and understand factors
that can complicate the interpretation of scarp dating techniques.
We examined fault scarps near the Dragon’s Back Pressure
Ridges in the Carrizo Plain, CA. The structure includes a system
of normal faults, which developed in the Pleistocene sedimentary
unit on the Northeast side of the San Andreas Fault. In March
2016, we used a Phantom-3 Drone with a GoPro camera and
took 195 aerial images of a graben on the northern extent of the
Dragon’s Back. The graben of interest strikes northwest and lies
on a regional slope of about 1° to the southwest. We ortho-
rectified and stitched together the drone images using Agisoft,
and generated a 1m-resolution DEM through Cloud Compare.
The tilt and mean height of the drone DEM were then corrected
using LiDAR data acquired in 2005 by NCALM, accessed
through Open Topography. We extracted 50 profiles from the
fault on each side of the graben, and analyzed the topographic
profiles using scarp diffusion models adjusted by a Gaussian
function with constants. This function estimates three model
parameters: the regional slope, the scarp height, and the
degradation coefficient (kt value) which represents relative
geomorphic ages (Hanks et al, 1984; Nash et al, 1984). We found
that the scarp downslope (southern scarp) had a lower kt value,
which a mean value of about 4.7m2, than the scarp upslope
(northern scarp), which had a mean value of 9.4m2. These
different geomorphic ages could be caused by several factors
including the relative orientation of the scarp to the regional
slope, vegetation density, and different formation ages. Our
study showed that surface processes influenced by regional
setting and local climate can produce apparently different
geomorphic ages of fault scarps according to current fault
diffusion models.
Low wave speed zones in the crust beneath SE Tibet
revealed by ambient noise adjoint tomography, Min Chen,
Hui Huang, Huajian Yao, Rob van der Hilst, and Fenglin Niu
(Poster Presentation 246)
We present a refined 3D crustal model beneath SE Tibet from
ambient noise adjoint tomography. Different from ray-theory-
based tomography, adjoint tomography in this study incorporates
a spectral-element method (SEM) and takes empirical Green’s
functions (EGFs) of Rayleigh waves from ambient noise
interferometry as the direct observation. The frequency-
dependent traveltime misfits between SEM synthetic Green’s
functions and EGFs are minimized with a preconditioned
conjugate gradient method, meanwhile the 3D model gets
improved iteratively utilizing 3D finite-frequency kernels. The new
model shows 3 – 6% shear wave speed increasing beneath the
western Sichuan Basin (SCB) (depth>15 km) and the central
Chuan-Dian Block (CDB), and 6 – 12% shear wave speed
reduction in the mid-lower crust beneath the northern and the
southern CDB. The inferred spatial pattern of low wave speed
zones, consistent with possible partial melt, suggests more
complex and disconnected geometry than the pervasive narrow
zone from the channel flow models.
Data-driven ambient noise correlation for characterization
and fragility evaluation in power grids, Qianli Chen, and
Ahmed E Elbanna (Poster Presentation 182)
The response of ambient noise has been used to cross correlate
and extract the Green’s function for general hyperbolic PDE’s in
many fields including acoustics, seismology and
helioseismology. The equipartitioning of wave field enable the
successful retrieval of the Green’s function of system with wave-
like solution. Recently, researchers found out that the electrical
power network consisting of transmission lines, generators, and
loads can be described by the propagation of electromechanical
(EM) waves in terms of the flow of power load and the inertia of
the synchronous generators. Utilizing the method of retrieval the
Green’s function from cross correlation and the approach of PDE
description of power grid system, we can extract the Green’s
functions for pairs of electrical generators and predict the
WELCOME
2016 SCEC Annual Meeting page 166
influence of potential disturbance from one power station on the
others. We apply this method first to a synthetic but realistic
power gird ambient frequency data using IEEE standard test
cases and real grid models for EI and WECC systems available
at UIUC. After comparing the cross-correlation with the Green’s
function and validating the results, we apply the method to real
PMU datasets. By building up this framework, we are able to
continue further work in the area of synchrophasor data analytics
as well as enhance the resilience improvements against cyber-
physical attacks.
Seismicity and spectral analysis in Salton Sea
Geothermal Field, Yifang Cheng, and Xiaowei Chen (Poster
Presentation 184)
The surge of “man-made” earthquakes in recent years has led to
considerable concerns about the associated hazards. Improved
monitoring of small earthquakes would significantly help
understand such phenomena and the underlying physical
mechanisms. In the Salton Sea Geothermal field in southern
California, open access of a local borehole network provides a
unique opportunity to better understand the seismicity
characteristics, the related earthquake hazards, and the
relationship with the geothermal system, tectonic faulting and
other physical conditions. We obtain high-resolution earthquake
locations in the Salton Sea Geothermal Field, analyze
characteristics of spatiotemporal isolated earthquake clusters,
magnitude-frequency distributions and spatial variation of stress
drops. The analysis reveals spatial coherent distributions of
different types of clustering, b-value distributions, and stress drop
distribution. The mixture type clusters (short-duration rapid
bursts with high aftershock productivity) are predominately
located within active geothermal field that correlate with high b-
value, low stress drop microearthquake clouds, while regular
aftershock sequences and swarms are distributed throughout the
study area. The differences between earthquakes inside and
outside of geothermal operation field suggest a possible way to
distinguish directly induced seismicity due to energy operation
versus typical seismic slip driven sequences. The spatial
coherent b-value distribution enables in-situ estimation of
probabilities for M≥3 earthquakes, and shows that the high large-
magnitude-event (LME) probability zones with high stress drop
are likely associated with tectonic faulting. The high stress drop
in shallow (1-3 km) depth indicates the existence of active faults,
while low stress drops near injection wells likely corresponds to
the seismic response to fluid injection. I interpret the spatial
variation of seismicity and source characteristics as the result of
fluid circulation, the fracture network, and tectonic faulting.
Multi-Scale Flash Heating and Frictional Weakening at
Seismic Slip-Rates in Rock, Frederick M Chester, Omid
Saber, and J. L. Alvarado (Poster Presentation 031)
Sliding rock-samples at rates > 1 mm.s^-1 can lead to a
significant reduction of friction, often referred to as dynamic
weakening. Thermal effects are understood to be the cause of
dynamic weakening, and the Flash Weakening (FW) mechanism
that can be activated at the initial stages of seismic slip by Flash
Heating (FH) at micrometer-scale contacts. FW can play a key
role in transient friction weakening leading to nucleation and
propagation of earthquake ruptures. We document FW in granite
(Westerly) by conducting sliding experiments at 1-1000 mm/s,
normal stress of 1-25 MPa, and displacement up to 40 mm using
a double-direct shear configuration with a 75 cm^2 sliding-
surface area. We document the characteristics of FH by direct
measurement of the flash-temperature distribution on sliding
surfaces using a high-speed Infra-Red (IR) camera. Images with
a resolution of 75 µm are captured at 300 frame/sec (every ~2.5
mm slip during high speed sliding) to document FH during sliding
and the subsequent cooling upon cessation of frictional slip.
Stepping velocity from quasi-static slip-rates to sustained high
velocities with high accelerations (up to 100g) over short slip
displacements (as low as 1 mm) allows us to observe the initial
stages of evolution of the transient friction. The IR images of the
sliding surface document a non-uniform distribution of
temperature generated by FH at the millimetric scale, reflecting
localization of normal and shear stresses to small portions of the
contact surface. Temperatures up to 550 C are observed on mm-
sized contact areas after ~10 mm of slip at ~800 mm/s sliding-
rate and moderate normal stress, while the majority of the
surface area displays significantly lower temperature. The
observations of FH are used to constrain models of flash
weakening that include the effects of macroscopic normal stress,
slip history and presence of heated contacts at multiple scales
(micrometer and millimeter). The models can explain the
evolution of transient friction with slip in the acceleration and
deceleration phases (e.g. the enhancement of FH with slip, and
hysteresis loops in the friction-velocity curve during acceleration
and deceleration) that is essential for modeling rupture
propagation and arrest.
SeedMe: Data sharing building blocks, Amit Chourasia,
David R Nadeau, John Moreland, Dmitry Mishin, and Michael L
Norman (Poster Presentation 348)
Computational simulations have become an indispensible tool in
a wide variety of science and engineering investigations. Nearly
all scientific computation and analyses create important transient
data and preliminary results. These transient data include
information dumped while a job is running, such as coarse output
and run statistics. Preliminary results include data output by a
running or finished job that needs to be quickly processed to get
a view of the job’s success or failure. These job output data
provide vital guidance that helps scientists review a current job
and adjust parameters for the next job to run. Quick and effective
assessments of these data are necessary for efficient use of the
computation resources, but this is complicated when a large
collaborating team is geographically dispersed and/or some
team members do not have direct access to the computation
resource and output data. Current methods for sharing and
assessing transient data and preliminary results are
cumbersome, labor intensive, and largely unsupported by useful
tools and procedures. Each research team is forced to create
their own scripts and ad hoc procedures to push data from
system to system, and user to user, and to make quick plots,
images, and videos to guide the next step in their research.
These custom efforts often rely on email, ftp, and scp, despite
the ubiquity of much more flexible dynamic web-based
technologies and the impressive display and interaction abilities
of today’s mobile devices. Better tools, building blocks, and
cyberinfrastructure are needed to better support transient data
and preliminary results sharing for collaborating computational
science teams.
WELCOME
2016 SCEC Annual Meeting page 167
SeedMe project is developing web-based building blocks and
cyberinfrastructure to enable easy sharing and streaming of
transient data and preliminary results from computing resources
to a variety of platforms, from mobile devices to workstations, and
make it possible to quickly and conveniently view and assess
results and provide an essential missing components in High
Performance Computing (HPC) and cloud computing
infrastructure. This work is an evolution of the SeedMe project [1,
2, 3] that will ultimately offer modular and flexible data sharing
building blocks to the computation community. The building
blocks will include authentication/authorization, granular access
controls, data sharing and indexing, micro format ingestion and
presentation for dashboard like functionality.
SeedMe building blocks is broadly applicable to a diverse set of
scientific and engineering communities, and SeedMe will be
released as a suite of open source modular building blocks that
may be extended by others. With this poster we'd like to
showcase the current progress on the project and engage with
the HPC community to get feedback.
Testing the effect of deficient real-time earthquake
catalogs on non-Poissonian earthquake likelihood
models: Examples from the Canterbury earthquake
sequence, Annemarie Christophersen, David A Rhoades,
David S Harte, and Matthew C Gerstenberger (Poster
Presentation 308)
The still on-going Canterbury earthquake sequence has provided
us with a wealth of seismological data. Among them are a series
of near-real time earthquake catalogs that show the deficiency of
real-time data. For example, while the final GeoNet catalog lists
20 M≥5.0 earthquake within the first 24 hours of the M7.1 Darfield
earthquake, the initial earthquake catalog had only 6 M≥5.0 in
the same time period. It took more than 18 months for the data
to be finalised. Other changes in the final catalog are relocation
of hypocenter and changes in earthquake magnitudes. We are
interested in the effect of the deficient real-time data on
earthquake forecasts. We calculated forecasts of durations
varying from 1 – 365 days from different implementations of the
Epidemic-Type-Aftershock-Sequence (ETAS) model and the
Short-Term-Earthquake-Probability (STEP) model. Some
models produce CSEP-type grid forecasts, others simulated
catalogs, and some both of these forms of forecasts. We employ
different testing methods to evaluate the effect of the changing
input data on the model forecasts and performance while
avoiding the Poisson assumption of the standard CSEP
consistency tests.
Deformation of the Xishancun Landslide, Sichuan,
inferred from seismicity, Risheng Chu (Poster Presentation
260)
Unstability of landslides, e. g. the slope movement and the
internal fracturing of the rock mass, can often generate
microseismicity, which is recorded as seismic signals on
seismographs. The temporal and spatial distribution of the
unstable regions is key to understanding the deformation of the
landslides and provides important information for landslide early
warning, which can be used as a supplementary to geodetic
monitoring of landslide surface. Toward this end, we deployed 30
seismometers on Xishancun landslide with a station spacing of
200~500 meters between the August and November of 2015.
The Xishancun landslide is about 80 million cubic meters and
located about 60 km northwest of the hypocenter of the 2008
Wenchuan earthquake. The landslide keeps deforming at about
5~8 cm/year after the earthquake, which poses great hazards.
Based on travel time and waveform characteristics, we found
many seismic events associated with the landslide and used their
waveform envelopes as templates. We then applied the sliding-
window cross-correlation technique to detect about 100 more
events. About 80% of the events are located on the foot of the
landslide while the others located on the head. The head and foot
of the landslide accumulate considerable deformation that
radiated seismic energy could be recorded by all stations. In
addition, there are many smaller events occurred in the main
body of the landslide whose released less energy can only be
recorded by a few stations. The distribution of the seismic events
may be related to the internal fractures and movement of the
landslide, which agrees with deformed areas monitored by
geodetic methods.
Localization and delocalization of shear in fault gouge
from thermal pressurization, Shanna Chu, and Eric M
Dunham (Poster Presentation 046)
Field observations of extreme shear localization in fault gouge
have been attributed to a number of factors, including thermal
pressurization. The extent of the localization is affected by
various parameters such as diffusion, friction, and gouge width.
A model of shear localization due to thermal pressurization
developed by Rice et al. (2014) and Platt et al. (2014) imposed a
constant slip velocity and solved for the evolution of shear
strength. In contrast, we couple the shear zone model to a simple
model for elasticity, and solve for the simultaneous evolution of
slip velocity and shear strength. This permits examination of not
only localization during rapid sliding but also delocalization as the
fault decelerates.
We are working on doing complete earthquake cycles in the
context of a spring-slider model to examine how shear localizes
and delocalizes over the full cycle.
Ground Motion Amplitude Variations in the Los Angeles
Region Measured by the Community Seismic Network,
Robert W Clayton, Robert Sanchez, Julian Bunn, and Richard
Guy (Poster Presentation 209)
We show variations in ground motion amplitude in the Pasadena
and northern Los Angeles regions due to the Borrego-Springs
earthquake (2016/06/10, M5.2) and other earthquakes in 2015-
2016. The measurements are from the Community Seismic
Network stations, which are deployed in private homes and
businesses as well as 100 LAUSD (Los Angeles Unified School
District) schools. The results show the enhanced shaking levels
due to the Los Angeles Basin and other smaller structures. Time
snapshots of the motions from Borrego-Springs shows that the
energy in downtown LA arrives both through the Los Angeles
Basin, and the San Gabriel/San Bernardino Basins at
approximately the same time, which accounts for some of the
amplification.
ShakeAlert Testing and Certification Platform: Point
Source and Ground Motion Based Evaluations, Elizabeth
S Cochran, Monica D Kohler, Douglas D Given, Jennifer
WELCOME
2016 SCEC Annual Meeting page 168
Andrews, Men-Andrin Meier, Egill Hauksson, Sarah E Minson,
Mohammad Ahmad, Jonathan DeLeon, and Stephan Guiwits
(Poster Presentation 187)
An earthquake early warning system, ShakeAlert, has been
developed over the last decade through a collaboration between
the U.S. Geological Survey and its partners. ShakeAlert
messages are composed of point source information generated
by three seismic-data based algorithms. In Spring 2016,
ShakeAlert transitioned from demonstration to production
prototype in California, with a similar switch in the Pacific
Northwest scheduled Fall 2016. The production prototype
system incorporates several features that improve the
robustness of the system, including a Testing and Certification
Platform. The Testing and Certification Platform provides a
quantitative measure of performance for evaluating modifications
to existing algorithms and assessing new algorithms. The
platform incorporates an in situ, real-time test and playback of a
common suite of test events. The test suite includes almost 100
different test cases: historic earthquakes in northern and
southern California, teleseismic events, and calibration and mass
recentering records.
We developed a set of metrics to evaluate the algorithm solutions
from the output of the test suite. We produce a count of matched,
missed, and false alerts by assigning threshold values for the
maximum allowable difference from the catalog value for
magnitude (2.0 units), epicenter (100 km), and origin time (15 s).
Additionally, we define a maximum time to first alert based on
when the alert may no longer be considered useful, i.e. an alert
is issued after S waves travel to a radius defined by a Modified
Mercalli Intensity of IV. An overall performance measure is
computed based on alert latency as well as accuracy of point
source parameters for magnitude, epicenter, and origin time.
ShakeAlert will soon incorporate finite-fault algorithms that
cannot be assessed using point source metrics. So, we are now
developing metrics based on how accurately an algorithm
predicts expected ground motion. To evaluate the current
algorithms, we take the point source solution and apply a set of
Ground Motion Prediction Equations (GMPEs) to estimate the
peak ground acceleration (PGA) at a given location, which are
then compared to those measured at seismic stations. For a
given PGA threshold we categorize the results into: true positive
(ground motions were correctly predicted to be above a given
threshold), true negative (ground motions were correctly
predicted to be below a given threshold), false (ground motions
were incorrectly predicted to be above a given threshold), and
missed (ground motions were incorrectly predicted to be below a
given threshold). Next steps will include finalizing the GMPE, or
suite of GMPEs, to be used in the assessments and adding a
temporal component that evaluates how algorithms perform as
an event evolves.
Biomarkers as a tool to measure coseismic temperature
rise, Genevieve L Coffey, Heather M Savage, Pratigya J
Polissar, Brett M Carpenter, and Cristiano Collettini (Poster
Presentation 025)
During earthquake slip, frictional resistance within a fault can
lead to the generation of extremely high temperatures. As a
consequence, investigating temperature rise within fault zones
provides a promising mechanism for the detection of past large
earthquakes. Evidence of frictional heating is often sparse, but a
variety of techniques have been successfully applied to identify
past heating events, these include vitrinite reflectance,
pseudotachylytes, and fission track thermochronology.
Here we explore the use of biomarkers as an alternative
approach to identification and quantification of temperature rise
in faults and apply it to different faults.
Biomarkers offer significant advantages as a tool to measure
temperature rise as they are abundant in the rock record,
undergo systematic structural changes as they are heated, and
are stable over a range of different time-temperature windows.
Specifically, we are interested in methylphenanthrenes, a
polycyclic aromatic hydrocarbon that is formed during the
diagenesis of organic material. As temperature increases we see
that the abundance of thermally stable methylphenanthrene
isomers also increases, while the abundance of thermally
unstable isomers decreases. Thermal maturity is quantified using
the modified methylphenanthrene index (MPI-3), which
increases with increasing maturity and is dependent both on the
temperature and duration of heating.
We present preliminary results from the Muddy Mountain thrust
in Nevada and the Spoleto thrust in the Northern Apennines of
Italy, both locations which demonstrate clear fault zone
architecture with well-defined principal slip zones (PSZ). High-
resolution sampling of these locations was completed to
construct temperature profiles across the fault zone and gain an
understanding of how small-scale complexity may influence
earthquake rupture.
Simulation of multi-segment kinematic rupture of the 1992
Landers earthquake and ground motion validation of
ground motion, Jorge G Crempien, and Ralph J Archuleta
(Poster Presentation 289)
The ability of earthquake rupture to jump across faults, or to
propagate on complex faults with multiple bends can significantly
affect the ground motion different sites might experience in the
near source region. Based on the dynamic rupture simulations of
the 1992 Mw7.2 Landers earthquake by Tago et al. (2012), we
have constructed a simple database of moment partitioning and
rupture initiation between the different segments. With this
information we have proceeded to simulate several kinematic
rupture realization using the UCSB method (Schmedes et al.,
2013; Crempien and Archuleta, 2015) for each of the three main
ruptured fault-segments: Camp Rock, Homestead and Johnson
Valley faults. The total moment is partitioned proportionally to the
area of each fault segment. We use the same moment correlation
length scaling law of Mai and Beroza (2002) for all three
segments.We simulate ground motion for all three fault-
segments and sum the contribution of each segment. The time
delay between the segments is based on the results of Tago et
al. (2012), though they are not necessarily exactly the same. We
compare the ground motion synthetics to actual recording using
different validation metrics reported by Goulet et al. (2015).
Preliminary results show almost no bias in response spectrum
with respect to actual recording. We also explore the bias with
randomization of the parameters related to moment partitioning
and rupture timing of each fault-segment.
WELCOME
2016 SCEC Annual Meeting page 169
Progress Report of the SCEC Utilization of Ground Motion
Simulations (UGMS) Committee, C.B. Crouse, and Thomas
H Jordan (Oral Presentation 9/14/16 09:00)
The goal of the UGMS committee, since its inception in the spring
of 2013, has been to develop digital long-period response
spectral acceleration maps for the Los Angeles region, for
inclusion in the NEHRP and ASCE 7 Seismic Provisions and in
the Los Angeles City Building Code. The maps are based on 3-
D numerical ground-motion simulations, generated using
CyberShake, and ground motions computed using the four NGA
West2 empirical ground-motion prediction equations that
incorporate a basin amplification term. The UGMS developed a
method to combine the response spectra from both approaches
to generate risk-targeted Maximum Considered Earthquake
(MCER) response spectra for the region. These response
spectra cover the period band from 0 to 10 sec and will be
proposed as an amendment to the ASCE -16 standard adopted
by the City of Los Angeles. The MCER response spectrum at a
given site would be obtained from a USGS-like web look-up tool.
Users would input the site coordinates and site class (or Vs30),
and the output would be a “site-specific” MCER response
spectrum and associated SDS and SD1 values of the Design
Response Spectrum. The user would have the option of
obtaining more information, such as hazard curves, risk
coefficients, and source and magnitude-distance deaggregation
data. The look-up tool is currently being developed by SCEC with
a scheduled release of a trial version toward the end of 2016 or
early 2017.
The work of the UGMS committee has been coordinated with (1)
the SCEC Ground Motion Simulation Validation Technical
Activity Group (GMSV-TAG), (2) other SCEC projects, such as
CyberShake and UCERF, and (3) the USGS national seismic
hazard mapping project.
Single station automated detection of transient
deformation in GPS time series with the relative strength
index: A case study of Cascadian slow-slip, Brendan W
Crowell, Yehuda Bock, and Zhen Liu (Poster Presentation 153)
The discovery of transient slow-slip events over the past decades
has changed our understanding of tectonic hazards and the
earthquake cycle. Proper geodetic characterization of slow-slip
events is necessary for studies of regional interseismic,
coseismic and postseismic tectonics, and miscalculations can
affect our understanding of the regional stress field and tectonic
hazard. Because of the proliferation of GPS data over the last
two decades, an automated algorithm is required to analyze the
signals and model the deformation on a station by station basis.
Using the relative strength index (RSI), a financial momentum
oscillator, we test the ability to detect events of various sizes and
durations. We first determine the statistics of the RSI under
different noise conditions and then use this information as the
basis for the automated transient detection algorithm by testing
different synthetic signals. We then apply the technique to daily
GPS displacement time series from 213 stations along the
Cascadia subduction zone to form a record of transient
deformation between 2005 and 2016. Our estimates of the size,
duration and propagation of major Episodic Tremor and Slip
(ETS) events are consistent with previous studies. We use the
detections to remodel the displacement time series and obtain
transient deformation rates over the past decade and discuss the
tectonic implications. Finally, we analyze the correlation between
transient detections and tremor showing good agreement
between the two at slab depths commonly associated with ETS
events.
Recent Improvements to AWP-ODC-GPU supported by
Blue Waters PAID Program, Yifeng Cui, Daniel Roten, Peng
Wang, Dawei Mu, Carl Pearson, Kyle B Withers, William H
Savran, Kim B Olsen, Steven M Day, and Wen-Mei Hwu (Poster
Presentation 344)
Supported by the NSF’s Petascale Application Improvement
Discovery (PAID) program, the SCEC computational team has
recently worked with the Improvement Method Enablers (IME)
team of UIUC, led by Wen-Mei Hwu, to optimize the nonlinear
AWP-ODC-GPU code for scalability and efficiency. During the
course of this project, plasticity and frequency-dependent
attenuation features have been introduced. This code integration
and optimization made it possible to run a nonlinear M 7.7
earthquake simulation on the southern San Andreas fault with a
spatial resolution of 25m for frequencies up to 4 Hz on 4,200 Blue
Waters GPUs at the National Center for Supercomputing
Applications.
Various strategies have been employed to optimize performance
and utilize accelerator technology, including yield factor
interpolation, memory tuning to increase occupancy,
communication overlap using multi-streaming, and parallel I/O to
support concurrent source inputs. To ensure enough occupancy
to hide memory latency and reduce redundant accesses, we use
shared memory to cache velocity variables, and texture cache
combined with register queues to cache read-only earth model
parameters. After these optimizations, the performance of the
plasticity kernel improved by 45% and the performance of the
stress kernel improved by 20%. The plasticity kernel reached
75% of theoretical maximum DRAM throughput and an access
fraction of almost 1.
By utilizing the multi-streaming technique, the linear
implementation of AWP-ODC-GPU is able to overlap
communication with computation, with a parallel efficiency of
100% achieved on 16,384 XK7 GPUs. Due to the larger
communicational volume required by the non-linear
implementation, the parallel efficiency initially dropped to 94.3%
on 3,680 Blue Waters XK7 nodes. We thus employed Cray’s
Topaware tool to aid node selection and MPI rank by mapping
the virtual Cartesian grid topology MPI layout onto the system's
torus interconnection network. Our nonlinear implementation
was able to obtain a speedup of 5.6%, and regain the parallel
efficiency to 99.5% at the scale. Overall, the computational cost
of the nonlinear code was reduced from initially 4x to less than
50% with respect to the highly efficient linear code, with 2.9
Pflop/s recorded in benchmarks on OLCF Titan.
Geology of liquefaction-induced lateral spreading, new
results from Christchurch, New Zealand, Gregory P De
Pascale, Jeffrey L Bachhuber, Ellen Rathje, Jing Hu, Peter
Almond, Christian Ruegg, and Mike Finnemore (Poster
Presentation 084)
Earthquake triggered liquefaction and lateral spreading was
widespread in Canterbury, New Zealand during the 2010-2012
WELCOME
2016 SCEC Annual Meeting page 170
Canterbury Earthquake Sequence (CES) and led to ~$20 Billion
NZD of damage to infrastructure. Permanent lateral ground
displacements and subsequent vertical subsidence along rivers
and streams progressed hundreds of metres back from the river
banks, significantly exceeding expected lateral spread distances
calculated from empirically-based predictive relationships.
Although the causes and timing of liquefaction during the CES
are well known, the geological controls on lateral spreading are
more poorly-understood. We undertook a multidisciplinary
investigation funded by the National Science Foundation (NSF)
that included trenching, mapping, satellite displacement models,
Multi-channel Analysis of Surface Waves (MASW), Cone
Penetration Testing (CPT), and boreholes in and around lateral
spreading sites in Christchurch - the main city in the Canterbury
region. The focus of the study is to better understand the geologic
controls on lateral spread occurrence and distances back from
river banks. Preliminary results based on the first lateral spread
trench (1 of 2) in our study demonstrate that: 1) displacements
observed from satellite-image modeling correspond well with on-
the-ground and subsurface displacement measurements and
headscarp margins; 2) trenching of Holocene sediments along
the headscarp margins of these lateral spreading zones
demonstrates that ground cracks and vertical steps at the surface
correspond with liquefaction-ejecta filled cracks in the subsurface
and a zone of extension indicated by normal faulting and horst
and graben structures; 3) translated sediments are composed of
laminar, sub-horizontal bedded sands with occasional silt
interbeds; 4) discrete displacements are well exhibited by
progressive vertical-lateral displacements of silt marker beds in
the sand and can be used to measure total movement of the
lateral spread mass; 5) radiocarbon dating indicates that the
translated sands are late-Holocene (<600 years BP); and 6)
displacements at the surface (from the CES) are less than the
cumulative displacement of the sand layers measured in the
subsurface (trench walls) which may suggest paleo-lateral
spreading at the site and therefore recurrence of lateral
spreading events. The preliminary results suggest repeated
lateral spreading (pre-historic and CES) in essentially the same
zones, and which we have found to have similar geologic
conditions. These results can provide a sound basis for hazard
zonation maps (e.g. liquefaction and lateral spreading), hazard
maps) and mitigation strategies (e.g. avoidance or ground
improvement).
Utilization of earthquake ground motions for nonlinear
analysis and design of tall buildings, Gregory Deierlein,
Nenad Bijelic, and Ting Lin (Oral Presentation 9/14/16 08:30)
One of the promising applications of simulated ground motions is
in the design of tall buildings and other unique structures, where
nonlinear dynamic analyses are used to evaluate their seismic
performance. In contrast to current practice, which requires
amplitude scaling of recorded ground motions to represent
extreme earthquake hazards, earthquake simulations can enable
more direct assessment without ground motion scaling or
reliance on empirical ground motion prediction equations.
Moreover, comprehensive simulations, such as SCEC’s
Cybershake study, hold the potential to more realistically capture
the influence of geologic basins and other geologic features on
ground motion intensities, frequency content, and durations. This
presentation will briefly review recent research on the use of
nonlinear structural analysis in design, with particular focus on
characterizing earthquake ground motions and their effects on
structures. The important influence of response spectral shape
and ground motion duration will be illustrated in a case study of
a twenty-story building. Efforts to compare and contrast the
response of structures to simulated and record motions will be
presented as a step towards evaluating and validating
earthquake simulations for use in engineering practice.
Geodetic Measurements of Slow Slip and Tremor in
Parkfield, CA, Brent G Delbridge, Roland Bürgmann, and
Robert M Nadeau (Poster Presentation 065)
It has been proposed that large bursts of deep tremor ( >20km
depth) near Parkfield, CA are associated with quasi-periodic
shear dislocations on the deep extent of the San Andreas Fault.
Geodetic studies have shown that slow slip accompanies tremor
in several subduction zones. However, prior to this study
deformation associated with tremor in a transform fault
environment had not been observed despite the ubiquitous
presence of tremor and LFEs and targeted attempts to observe
this deformation. In this study we report geodetic measurements
of surface strains associated with large tremor swarms that are
inferred to be concurrent with slow-slip events with moment
magnitudes exceeding 5. The strain rates associated with these
events are below the detection level of GPS networks, thus in
order to observe this deformation we have utilized two long-
baseline laser strainmeters (LSM) located in Cholame, CA. In
order to overcome a small signal-to noise-ratio in the strainmeter
data, we have stacked the strain records associated with more
than 50 large tremor-burst events, each approximately 10 days
in duration. The average surface strains associated with these
events are on the order of several nanometers and correspond
to fault slip on the order of 5 millimeters per event (assuming a
fault patch extending ~25 km along-strike and ~15km in depth).
The measured moment associated with these events is a factor
of two smaller than previously proposed based on theoretical
estimates.
Ground motion amplification in the Kanto Basin from
future Itoigawa-Shizuoka earthquakes near Tokyo using
virtual earthquakes., Marine A Denolle, Pierre Boué, Naoshi
Hirata, Shigeki Nakagawa, and Gregory C Beroza (Poster
Presentation 273)
The Tokyo Metropolitan area is subject to high seismic risk due
to the nearby triple junction and because it is sitting atop the deep
Kanto sedimentary basin. In addition, the Itoigawa-Shizuoka
Tectonic Line (ISTL) is one of the crustal faults that comprises a
zone of active plate boundary, can host oblique-strike slip
earthquakes with magnitudes greater than 7, with up to 30% of
probability in 30 years (The Headquarter of Earthquake
Research Promotion, 2015), and is only ~ 150-200km away from
Tokyo. Currently at a mature stage in its seismic cycle, predicting
strong ground motion from earthquakes on it is important for
seismic hazard assessment. Previous studies using the ambient
field have suggested that ISTL sources might have particularly
strong amplification in the Kanto Basin (Denolle et al., 2014). We
take advantage of the dense seismic network MeSO-net
(MEtropolitan Seismic Observation network) and of the locations
of the Hi-net (High Sensitivity Seismograph network) stations
near the ISTL to build virtual earthquakes from the ambient
WELCOME
2016 SCEC Annual Meeting page 171
seismic field and predict strong ground motion in the 1 – 10
seconds period range.
First, we construct the Earth’s impulse responses from
deconvolution of the ambient seismic field recorded between Hi-
net stations (as sources) and MeSO-net stations (as receivers)
with a method that preserves the relative amplitudes of the
response. Then, we correct for ray bending by optimally rotating
the 9 component Green tensor to minimize the coupling between
transverse and non-transverse components. Sharp edges
around the basin are responsible for scattering of fundamental-
to-first overtone surface-wave mode, with magnified amplitude
and extended shaking duration. We apply the virtual earthquake
approach that turns each station-source near the fault into a
double-couple point source by correcting the Green tensor for
source depth and mechanism. The depth and mechanism
correction is a source of uncertainty in peak ground motion and
in shaking duration, which we quantify using simple velocity
models. To verify that our relative amplitudes are reliable and to
calibrate them to absolute levels, we use recordings of the M6.7
Northern Nagano earthquake that occurred on November 22,
2014 at the northern end of the fault segment to calibrate
absolute amplitudes. Finally, we estimate the long period ground
motions for a suite of of M7+ scenario ruptures on the ISTL and
explore the role of waveguides, basin resonance, and basin
edges in strong ground motion.
Surface slip rate of the Imperial Fault estimated from
remote controlled quadcopter photogrammetry, John
DeSanto, and David T Sandwell (Poster Presentation 134)
In recent years, advances in photogrammetry have allowed
remote controlled quadcopters to emerge as a useful tool for
remote geological surveying. These tools allow pilots to collect
photographic data of difficult to reach outcrops and create a
three-dimensional model for easy interpretation. However, the
geodetic applications of this technique are limited by the poor
accuracy of the quadcopter GPS, which introduces distortions
into generated 3D models. To minimize such distortions, we
couple quadcopter imagery with independent campaign GPS
measurements, which serve as ground truths. In 2016 we
conducted a campaign-style GPS Survey along the Imperial
Fault, occupying monuments adjacent to an intersection between
the fault trace and a concrete canal along which surface fault
displacement is readily visible. We conducted a concurrent
quadcopter survey of these monuments, collecting photographic
images of the canal as it crosses the fault trace. We process
these images using the Agisoft Photoscan Pro software to create
a three dimensional model of the canal. From this model, we
measure a displacement in the canal of 88-102 cm. Comparing
this displacement estimate to historical measurements made
following the October 15, 1979 Imperial Valley earthquake yields
a surface slip rate of 7-11 mm/yr, consistent with accepted
values.
USGS Southern California GPS Network, Daniel N
Determan, Aris G Aspiotes, Derik T Barseghian, Kenneth W
Hudnut, and Keith F Stark (Poster Presentation 145)
The USGS Pasadena field office now operates 140 permanent,
continuously-operating Global Positioning System monitoring
stations as part of the Southern California GPS Network (SCGN).
The SCGN network has grown and modernized significantly over
the last several years with the help of projects like the American
Recovery and Reinvestment Act (ARRA), which provided funding
to replace equipment throughout the network and added the
capability of tracking other Global Navigation Satellite Systems
(GNSS) signals, such as Russia’s GLObal NAvigation Satellite
System. Another project, funded by an Urban Area Security
Initiative (UASI) grant received by Caltech, provided resources
for the purchase of 41 Trimble NetR9 GPS receivers and Zephyr
Geodetic II antennas which have been added to the network.
This new equipment supports other GNSS and it includes “on-
board” Precise Point Positioning with Ambiguity Resolution
(PPPAR) processing. This new capability allows the streaming of
processed positions directly from the receiver, which is well-
suited to support the West Coast - Earthquake Early Warning
(WC-EEW) system at sites close to an active fault for determining
displacements quickly after a large event. The UASI project also
provided us the opportunity to co-locate GPS equipment at many
newly upgraded UASI seismic stations, expanding the network
from 104 to 140 stations, increasing the number of real-time
stations from 95 to 130, and increasing the number of co-located
seismic/GPS stations from 26 to more than 60. We process all
our real-time stations using RTNet software in precise point
position mode and we have partially implemented differential
processing at a subset of our sites. We will be including all
stations in our differential processing in the near future. The 41
Trimble NetR9 stations also compute on-board PPPAR positions
using RTX software. We use both the RTX and RTNet PPPAR
processing output to compare positions and for an independent
solution for quality control. Our network utilizes three types of
geodetic grade GPS receivers: 89 have NetG3A receivers; 10
utilize NetRS receivers; and 41 use the new NetR9 receivers with
RTX software (PPPAR). In the coming year, we hope to replace
the NetRS receivers and continue integrating with the Northern
California GPS Network and the Southern California Seismic
Network. We also plan to concentrate more on improving our
data telemetry, to increase the robustness and reliability of all of
our real-time data streams for EEW and other products.
Learning viscoelasticity with neural networks, Phoebe
DeVries, Thomas B Thompson, and Brendan J Meade (Poster
Presentation 018)
Viscoelastic models have been widely used to explain geodetic
observations of the earthquake cycle, time-dependent stress
transfer, and delayed earthquake triggering. The calculations
involved in these modeling efforts are often computationally
intensive; as a result, studies tend to adopt a few fixed
rheological structures and model geometries, and examine the
associated predictions of stress evolution over short (<10 yr) time
periods at a given depth or specific location after a large
earthquake. Training a deep neural network to accurately
approximate these viscoelastic solutions – at any time, location,
and for a large range of rheological structures – allows these
calculations to be done quickly, for geometrically complex faults
and over many earthquake cycles, with arbitrarily high spatial and
temporal resolution. Preliminary tests suggest this method could
accelerate these calculations by a factor of 500 and perhaps
orders of magnitude more; this kind of acceleration will facilitate
an understanding the potential viscoelastic effects of large
earthquakes across wider ranges of model parameters and at
larger spatial and temporal scales than have previously been
possible.
WELCOME
2016 SCEC Annual Meeting page 172
Progress towards a comprehensive incremental slip rate
and paleo-earthquake age and displacement record for
the central Garlock fault, James F Dolan, Sally F McGill, Ed J
Rhodes, and Thomas M Crane (Poster Presentation 101)
As part of our ongoing studies of the incremental slip-rate and
paleo-earthquake age and displacement record of the Garlock
fault, we present new slip rate results from our Pilot Knob Valley
(PKV) site along the east-central part of the fault. These
incremental rate data are based on geologic mapping and lidar
terrain analysis of several progressive offsets of fluvial terrace
risers that we have dated with the robust new post-IR IRSL
luminescence geochronometer. Our preliminary interpretation of
the new data is that they generally confirm our earlier incremental
slip rate record from farther west along the Garlock fault, with a
very rapid slip rate over the past ca. 2 ka preceded by a lull in
earthquake activity and an absence of slip for a several-
thousand-year-long period during the early late to late mid-
Holocene. Slightly different offset measurements between the
west-central (Christmas Canyon West & El Paso Peaks) sites
and the new east-central PKV site leave open the possibility that,
while the fault exhibited the same general pattern of millennia-
long, alternating fast and zero slip rate periods, the actual
number of earthquakes and their displacements may have been
different between the two areas, with the east-central site
perhaps experiencing more and/or larger-displacement events.
This suggestion needs to be confirmed with additional data,
however, and should be considered preliminary.
Access to Geodetic Imaging Products through
GeoGateway Analysis, Modeling, and Response Tools,
Andrea Donnellan, Jay W Parker, Robert A Granat, Michael B
Heflin, Marlon E Pierce, Jun Wang, John B Rundle, and Lisa
Grant Ludwig (Poster Presentation 343)
GeoGateway is a web map analysis, modeling, and response
tool to allow users to efficiently find and use NASA geodetic
imaging data products. GeoGateway tools steer users to relevant
products using automated feature extraction, manual search
tools, and data quality metrics. GeoGatway goals are to a)
simplify the discovery of geodetic imaging products; b) enable
researchers to explore and integrate data products through
online or offline analysis; and c) allow researchers to easily
share, publish, and collaborate on results of these online
experiments. GeoGateway increases the value of existing
geodetic imaging products to researchers through automated
machine learning and computer vision techniques to identify
features, anomalies, and artifacts bridging the gap between
production and end-use of data products. Individual data types
are far more powerful when they are used in conjunction with
other products to test a hypothesis. Automated and manual
approaches are be available for the user through the interface.
Users can perform time and geographically limited search for
products. Time windowing is particularly important because
crustal deformation velocities vary over time as earthquakes and
postseismic deformation occur, causing offsets and rate
changes. Statistics on GPS and UAVSAR data will enable users
to separate data artifacts and errors from geophysical signals.
Heterogeneous crustal deformation data come in the form of
GPS position time series and radar or optical maps of
components of surface deformation. Earthquakes and
postseismic response complicate the already heterogeneous
data by adding episodic jumps and changing rates of surface
deformation to the products. Data volumes will become even
larger and the surface deformation data more complicated when
NASA launches NISAR in late 2021. GeoGateway components
are open source and available for reuse by other projects and
researchers. The emphasis on facilitating routine workflow
promotes the reuse of techniques necessary to advance
earthquake research.
LArge-n Seismic Survey in Oklahoma (LASSO): Probing
injection-induced seismicity with a dense array, Sara L
Dougherty, Elizabeth S Cochran, and Rebecca M Harrington
(Poster Presentation 197)
In response to the increased seismicity in Oklahoma, we
deployed a temporary large-N array of more than 1,800 vertical-
component nodal seismometers over a 25-km-by-32-km region
(nominal spacing of ~400 m) in northern Oklahoma during spring
2016. This dense array will allow us to view sequences of likely-
induced seismicity in a region of active wastewater injection with
unprecedented clarity. We will use this LArge-n Seismic Survey
in Oklahoma (LASSO) array to assess the locations, frequency,
magnitudes, source properties, and spatiotemporal evolution of
micro- to minor earthquakes in an effort to improve our
understanding of the relationship(s) between injection
parameters and induced seismicity, potentially leading to more
accurate seismic hazard assessment and improved hazard
mitigation. We will also identify the locations and orientations of
subsurface faults to provide insights into where future locations
of injection-induced seismicity may occur and perform
tomographic imaging of the shallow crust to provide additional
details of the geologic structure and any potential preferential
location of seismicity in or along a particular unit or fault.
Calculating regional stresses for northern Canterbury: the
effect of the 2010 Darfield earthquake, Susan Ellis, Charles
A Williams, John Ristau, Martin Reyners, Donna Eberhart-
Phillips, and Laura M Wallace (Poster Presentation 021)
We model regional stresses before and after the Mw 7.1 Darfield
earthquake of September 2010 in Canterbury, New Zealand
including crustal structure derived from seismic tomography.
Models show that the Banks Peninsula volcanic assemblage acts
as a strong, rigid block that pinches out ductile layers in the mid-
crust but has little effect on shallower principal stress
orientations. Static stress changes from the Darfield earthquake
are everywhere <25 MPa except within 5 km of the fault. When
added to regional stresses, these create only small rotations of
<5° in the orientation of maximum horizontal stress SHmax, even
near the fault. Predicted stress rotations do not correlate strongly
with those inferred from aftershock focal mechanisms. The
perturbations caused by earthquake stresses are not significant
enough to explain either the magnitude or the sense of SHmax
rotations near the fault at seismogenic depths.
Applying newly developed luminescence dating to alluvial
fans in the Anza Borrego Desert, southern California,
Brittney Emmons, Seulgi Moon, Nathan D Brown, Kimberly D
Blisniuk, and Ed J Rhodes (Poster Presentation 117)
Over recent decades, multiple geochronologic tools have been
used to date various geomorphic offsets (e.g., alluvial fans, fluvial
terraces), enabling long-term slip rate studies of active faults. In
WELCOME
2016 SCEC Annual Meeting page 173
this study, we show the robustness of the newly developed ‘p-IR-
IRSL225’ technique as a geochronometer to date alluvial fan
deposits in the Anza Borrego Desert of southern California. This
new method is based on two infrared stimulated luminescence
(IRSL) measurements, one at 50°C and another at 225°C
(referred to as p-IR-IRSL225) and takes advantage of potassium
feldspar grains that have a brightness higher than quartz used in
OSL dating (Rhodes, 2015). We applied this technique to date
alluvial fans that were offset by the San Jacinto fault zone and
were previously dated with 10Be exposure of surface clasts and
U-series disequilibria from pedogenic carbonates from
subsurface clast rinds (Blisniuk et al., 2012). Initial results from
the p-IR-IRSL225 method without fading correction yield dates of
5.5 +/- 0.4 (1 sigma) ka in Ash Wash, and 6.5 +/- 0.7 ka and 5.2
+/- 0.7 ka in Santa Rosa Mountains sites in the Anza Borrego
Desert. These p-IR-IRSL225 dates are consistent with dates
from U-series of soil carbonates, but are slightly younger than
those from 10Be exposure at each location. Depositional dates
from p-IR-IRSL225 indeed lay between the maximum 10Be
exposure ages and the minimum U-series ages. These robust
results suggest that the p-IR-IRSL225 dating technique is an
excellent tool for quantifying the rate of fault movement where
other dating methods cannot be applied or where dual dating
techniques should be applied to extend the record of active fault
deformation in southern California and elsewhere around the
world.
Plasticity Throughout the Earthquake Cycle, Brittany A
Erickson, Eric M Dunham, and Jeremy E Kozdon (Poster
Presentation 045)
We are developing an earthquake cycle model to simulate
multiple ruptures in complex geometries, with material
heterogeneity and off-fault plastic response. This initial study
focuses on the 2D antiplane shear case of ruptures on strike-slip
faults with rate-and-state friction and off-fault plasticity. Both rate-
independent plasticity with hardening and viscoplasticity are
considered, where stresses are constrained by a Drucker-Prager
yield condition. The volume is discretized using a finite difference
method with interseismic loading applied by imposed motion of
the remote boundaries rather than through backslip. Quasi-
dynamic events nucleate at depth and propagate toward the free
surface, carrying stress perturbations that result in off-fault plastic
strain. With a depth-dependent yield stress, the magnitude and
extent of plastic strain is affected by cohesion, viscosity, and
isotropic hardening. For rate-independent plasticity with
hardening, the yield surface expands with plastic flow.
Consequently, the first rupture in the cycle causes plastic strain,
but for all subsequent events the off-fault material responds
elastically. For perfect viscoplasticity, all ruptures generate
plastic strain. If hardening is included in the viscoplastic model,
the first event causes the most plastic strain (in a ~100 m region
away from the fault at the free surface for most scenarios), with
a decreasing amount of additional plastic strain for each
subsequent rupture. In all cases we found that a shallow slip
deficit occurs with the first event, but that the evolution of the
deficit with each subsequent event is dictated by the plasticity
model. In particular, there is almost no change in the deficit when
rate-independent plasticity is used and a continuous increase
with the viscoplastic models. Integration of the off-fault plastic
strain from the viscoplastic model reveals that a significant
amount of offset is accommodated by inelastic deformation (~1
m per 10 ruptures). Motivated in part by these results, we are
developing a more complete cycle model that can account for
interseismic loading and dynamic rupture. To do this, we are
developing a unified code capable of simulating both quasi-static
and dynamic rupture problems. The dynamic rupture code is
based on a discontinuous Galerkin method, and will be capable
of handling complex geometries, heterogeneous material
properties, and utilizing dynamically adapted grids.
Geodetic slip rate estimates in California, and their
uncertainties, Eileen L Evans (Poster Presentation 147)
Current understanding of the seismic potential of faults in
California is limited in part by our ability to resolve spatial and
temporal changes in fault slip rates across the Pacific-North
American plate boundary, and quantify their uncertainties. Fault
slip rate can be estimated by modeling fault systems, based on
space geodetic measurements of surface ground displacement
(GPS and InSAR). However, models that include elastic
deformation due to locked faults require fault geometries to be
prescribed, and geodetic slip rate estimates may vary widely due
to measurement and epistemic (model) uncertainties. To
examine published geodetic slip rate estimates in California and
quantify variability among models, we compile 31 published
geodetic slip rate studies in California and Nevada. Because
deformation models may vary in the number of faults represented
and the precise location of faults, we combine published geodetic
slip rate estimates on a georeferenced grid and compare models
spatially. Within each grid cell, a number of metrics are
considered based on the suite of fault slip rates in the cell. These
metrics include geometric moment (potency), strain and rotation,
and variation among models. This approach assumes that all
published geodetic slip rate estimates are equally valid, and
therefore this variability among models serves as a proxy for
epistemic uncertainties in geodetic slip rates: we find an average
standard deviation in potency rate of 1.5×10^6 m^3⁄yr for cells of
725 km^2 (1,365 cells), which corresponds to ~2 mm/yr of model
uncertainty on a given slip rate. These uncertainties may be
incorporated into hazard estimates, enable rigorous comparison
with geologic slip rates, and used to systematically identify
regions that may require more careful consideration in terms of
modeling available geologic and geodetic data.
Differential Waveform Analysis to Test for a Mid-Crustal
Low-Velocity Zone Beneath the Western Mojave, William
K Eymold, Thomas H Jordan, and David A Okaya (Poster
Presentation 232)
Lee et al. (JGR, 2014a) have applied full-3D tomography to a
large set of three-component earthquake waveforms and
vertical-component ambient-field correlagrams to obtain a
revised community velocity model for Southern California, CVM-
S4.26. An interesting feature of this model is a distinct mid-crustal
low-velocity zone (LVZ) beneath the western Mojave region with
a minimum P velocity of 6.0 km/s and a minimum S velocity of
3.4 km/s at 14-km depth, which is overlain by a high-velocity zone
(HVZ) with maxima of 6.7 km/s and 3.9 km/s, respectively, at 6
km. Although CVM-S4.26 provides a good fit to three-component
seismograms for paths crossing this region (Lee et al., SRL,
2014b), the model is constrained to be isotropic. Geologic
reconstructions of Mojave-block development suggest that the
region could be underlain by highly anisotropic schists emplaced
WELCOME
2016 SCEC Annual Meeting page 174
during the transition from subduction to transform faulting. One
issue, therefore, is whether the HVZ-LVZ reversal is an artifact
resulting from the inversion of three-component waveform data
under an isotropic assumption. To test this hypothesis, we have
analyzed the frequency dispersion of single-component
waveforms for Mojave-crossing paths with sources at LVZ
depths. Using a 1D approximation, we show that a diagnostic
feature of the HVZ-LVZ structure is a reduction of the differential
phase delays between the S wave and surface wave at
frequencies tuned to the near-surface and mid-crustal
waveguides. We describe our preliminary attempts to verify the
existence of the mid-crustal LVZ in this region using the available
earthquake data.
Local near-instantaneously dynamically triggered
aftershocks of large earthquakes, Wenyuan Fan, and Peter
M Shearer (Poster Presentation 173)
Aftershocks are often triggered by static- and/or dynamic-stress
changes caused by mainshocks. The relative importance of the
two triggering mechanisms is controversial at near-to-
intermediate distances. We detect and locate 48 previously
unidentified large early aftershocks triggered by 7 <= M < 8
earthquakes within a few fault lengths (~300 km), during times
that high-amplitude surface waves arrive from the mainshock (<
200 s). The observations indicate that near-to-intermediate field
dynamic triggering commonly exists and fundamentally promotes
aftershock occurrence. The mainshocks and their nearby early
aftershocks are located at major subduction zones and
continental boundaries, and mainshocks with all types of faulting-
mechanisms (normal, reverse, and strike-slip) can trigger early
aftershocks.
Quantifying Late Quaternary deformation in the Santa
Maria Basin: A OSL, GPS and soil chronosequence
based model for determining strath terrace deformation in
the Zaca Creek drainage, Santa Barbara County, Andrew
C Farris, and Nate W Onderdonk (Poster Presentation 016)
The Santa Maria Basin is located on the Central Coast of
California and comprises the area bound by the Santa Ynez Fault
to the south, and the Little Pine-Foxen Canyon Fault Zone to the
north. The Santa Maria Basin is a zone of convergence between
the obliquely rotating Western Transverse Ranges and the non-
rotated Southern Coast Ranges. Although early Quaternary
shortening due to crustal convergence with the Western
Transverse Ranges is well documented in the Santa Maria Basin,
the style and amount of Late Quaternary deformation is unknown
in the region. 7 levels of fluvial terraces in the Zaca Creek
drainage offer significant insight into Late Quaternary
deformation. Soil test pits were used for correlative purposes in
a setting with a blind terrace stratigraphy and potentially offset
terrace structures. Each terrace tread’s soil morphology was
described according to standardized United States Department
of Agriculture soil survey methods. The oldest strath terrace soils
correlated with generally high levels of soil genesis, thick illuvial
horizons, and localized iron oxide induced hardpans due to low
soil pH. The youngest fill terraces correlated with generally lower
levels of soil genesis with localized lithologic discontinuities.
Detailed field mapping and surveying with a differential GPS unit
suggests significant deformation in the oldest terraces (terrace
levels Qt4 to Qt6), which are likely older than at least 40 ka.
Terrace warping occurs in two separate locations due to
Quaternary folding and convergence related regional shortening.
Select strath terrace treads were sampled for Optically
Stimulated Luminescence analysis for age control. After the OSL
based terrace chronology is complete, the style and rate of
deformation and regional uplift in the Zaca Creek area will be
able to be quantified. Results from this study will have significant
implications for seismic risk, regional tectonics, and soil-
landscape relationships in clay soil and agricultural settings.
Mento Carlo Inversion of the 1D velocity and anisotropic
model in the Juan de Fuca plate , Tian Feng, Thomas
Bodin, and Lingsen Meng (Poster Presentation 230)
In this study, we demonstrate the advantage of the Monte Carlo
method in solving 1D velocity models. In order to obtain accurate
models, both velocity and anisotropy need to be accounted for in
the inversion. However, the unknown number of layers of
isotropic and anisotropic velocity separated by discontinuities
makes it challenging to set up the model parameters. The
number of parameters could also alter during the Monte Carlo
search process, which controls the speed and resolution of
inversion. The core of the Markov Chain Mento Calro (MCMC)
inversion method is based on the Bayesian Law: p(m|d) = p(d|m)
x p(m). In this work, we derive the velocity and anisotropic
models from Rayleigh wave dispersion curves using the MCMC
method. We conducted intensive synthetic tests to verify the
accuracy of this method. We then apply this method to the
Rayleigh dispersion curves observed at the Juan de Fuca Plate
and obtain anisotropic 1D velocity models with reasonably high
resolution. In comparison with SKS splitting observations, the
overall mean spitting times predicted by the 1D model based on
Rayleigh waves are smaller and the fast axis more concentrated.
This phenomenon can be explained by the difference of path
effects between the SKS waves and Rayleigh waves: The SKS
phase samples the whole depth range from the core-mantle
boundary to the receiver station. On the other hand, Rayleigh
waves are mainly sensitive to the crustal and upper mantle
structures beneath the receiver station. In our future work, similar
analysis will be extended to other regions to potentially separate
the contributions of shallow and deep anisotropy in the SKS
splitting observations.
Physical controls of spontaneous and triggered slow-slip
and stick-slip at the fault gouge scale, Behrooz Ferdowsi,
and David L Goldsby (Poster Presentation 038)
Fault slip modes span a continuum of behaviors from tremors
and slow slips to earthquakes. Recent laboratory studies reveal
that a spectrum of slow-slip responses emerge near the
threshold between stable and unstable failure, governed by the
complex interplay of frictional properties, effective normal stress
and the elastic stiffness of the damaged and undamaged zones
of the fault. Here we use simulations of particle dynamics at the
micro- to meso-scales of fault gouge to explore these
observations. The virtual experiments are carried out at a range
of loading conditions, including confining pressures of 0.5 to 50
MPa and boundary shear rates of 10-6 to 10-1 m/s. Our granular
system consists of grains with the mechanical properties of
quartz. The grains interact with each other in the normal direction
via a Hertzian contact law and in the shear direction via a
WELCOME
2016 SCEC Annual Meeting page 175
Coulomb friction law. The shear velocity is applied through a
spring attached to the rigid top boundary of the gouge, with
values of the spring stiffness that vary from infinity (the stiff/rigid
spring limit) to finite, and soft, values. In the limit of infinite spring
stiffness, the dynamical behavior of the gouge transitions from
stick-slip to continuous (steady) sliding as the shear rate is
increased from low to high values. For a finite and soft spring
stiffness, the relative stiffnesses of the spring and gouge, and the
feedback between these two stiffnesses, set the internal gouge
dynamics and control the slip modes. We use grain-scale forces
and displacements obtained from numerical simulations to
reconstruct the variation of local stress versus local slip
displacement for the spectrum of slip modes. Previous
observations by Lehman et al. (Nat. Commun., 2016) suggest
that slow earthquakes are governed by the same instabilities in
dynamical phase space (stiffness, k, velocity, V and pressure, Pn
) as normal earthquakes; however, this hypothesis needs to be
tested at the local and grain scales. To accomplish this, we
measure local stresses, deformation components and
displacements in numerical simulations for different slip modes
to compare and contrast the dynamical transitions that give rise
to these modes. Our ultimate goal is to provide physically-based
friction laws with predictive power across the length scales that
can in turn describe the response of gouge-filled faults to static
and dynamic stresses in slow-slip and fast-slip regimes.
A Spatiotemporal Clustering Model for the Third Uniform
California Earthquake Rupture Forecast (UCERF3-
ETAS) – Toward an Operational Earthquake Forecast,
Edward H Field, Kevin R Milner, Jeanne L Hardebeck, Morgan T
Page, Nicholas J van der Elst, Thomas H Jordan, Andrew J
Michael, Bruce E Shaw, and Maximilian J Werner (Poster
Presentation 298)
We, the ongoing Working Group on California Earthquake
Probabilities, present a spatiotemporal clustering model for the
third Uniform California Earthquake Rupture Forecast
(UCERF3), with the goal being to represent aftershocks, induced
seismicity, and otherwise triggered events as a potential basis for
Operational Earthquake Forecasting (OEF). Specifically, we add
an Epidemic Type Aftershock Sequence (ETAS) component to
the previously published time-independent and long-term time-
dependent forecasts. This combined model, referred to as
UCERF3-ETAS, collectively represents a relaxation of
segmentation assumptions, the inclusion of multi-fault ruptures,
an elastic-rebound model for fault-based ruptures, and a state-
of-the-art spatiotemporal clustering component. It also
represents an attempt to merge fault-based forecasts with
statistical seismology models, such that information on fault
proximity, activity rate, and time since last event are considered
in OEF. We describe several unanticipated challenges that were
encountered, including a need for elastic rebound and
characteristic magnitude-frequency distributions on faults, both
of which are required to get realistic triggering behavior.
UCERF3-ETAS produces synthetic catalogs of M≥2.5 events,
conditioned on any prior M≥2.5 events that are input to the
model. We evaluate results with respect to both long-term (1,000-
year) simulations, as well as for 10-year time periods following a
variety of hypothetical scenario main shocks. While the results
are very plausible, they are not always consistent with the simple
notion that triggering probabilities should be greater if a main
shock is located near a fault. Important factors include whether
the magnitude-frequency distributions near faults includes a
significant characteristic earthquake component, as well as
whether large triggered events can nucleate from within the
rupture zone of the main shock. Because UCERF3-ETAS has
many sources of uncertainty, as will any subsequent version or
competing model, potential usefulness needs to be considered
in the context of actual applications.
The mechanics of multifault ruptures and the keystone
fault hypothesis, John M Fletcher, Michael E Oskin, and
Orlando J Teran (Poster Presentation 074)
Regardless of global tectonic regime, most large earthquakes
activate slip on more than one fault. Likewise, the magnitude of
an earthquake increases substantially with the number of faults
that become activated. Despite the importance of multifault
ruptures for forecasting seismic hazard, their genesis remains
poorly understood, and classical applications of static yield
criteria are inadequate to describe the mechanical conditions
required to prepare multiple faults with diverse orientations to fail
simultaneously in a single earthquake. This is because the critical
stress level for fault failure depends greatly on fault orientation
and is lowest for optimally oriented faults positioned
approximately 30° to the greatest principal compressive stress.
Yet, misoriented faults whose positioning is not conducive to
rupture are also commonly activated in large earthquakes.
The 2010 El Mayor-Cucapah earthquake of magnitude Mw 7.2
propagated through a network of high- and low-angle faults
producing the most complex rupture ever documented on the
Pacific-North American plate margin. Our extensive database of
mapped fault scarps and offset geomorphic markers
demonstrate systematic changes in coseismic slip direction with
fault orientation. Using stress inversions of surface displacement
and seismic data, we find that the El Mayor-Cucapah earthquake
initiated on a fault, which due to its orientation, was among those
that required the greatest stress for failure. Although other
optimally-oriented faults must have reached critical stress earlier
in the interseismic period, Coulomb stress modeling shows that
slip on these faults was initially muted because they were pinned,
held in place by misoriented faults that helped regulate their slip.
In this way, faults of diverse orientations could be maintained at
critical stress without destabilizing the network. We propose that
regional stress build-up continues until a misoriented keystone
fault reaches its threshold and its failure then spreads
spontaneously across the network in a large earthquake. In
addition to explaining the mechanics of multifault ruptures, our
keystone fault hypothesis provides new understanding for the
seismogenic failure of severely misoriented faults like the San
Andreas Fault and the entire class of low-angle normal faults.
Sedimentary provenance constraints on the Quaternary
faulting history of the Mission Creek fault strand, southern
San Andreas Fault Zone, CA, Julie C Fosdick, Kimberly D
Blisniuk, and Louis Wersan (Poster Presentation 122)
Quaternary alluvial fan deposits along the foothills of the San
Bernardino Mountains record sediment dispersal from upland
catchments across the Mission Creek fault strand of San
Andreas fault in the San Gorgonio Pass, CA, and thereby yield
key markers of the fault’s lateral and vertical displacement
WELCOME
2016 SCEC Annual Meeting page 176
history. Here, we present an integrated provenance analysis of
conglomerate, sandstone, and detrital zircon collected from early
Pleistocene (?) Deformed gravels of Whitewater River (Qd), the
overlying flat-lying alluvial-fanglomerate (Qo, Cabezon
Fanglomerate), and modern drainages to evaluate catchment
sources within the San Bernardino Mountains (SBM) and Little
San Bernardino Mountains (LSBM). A statistical analysis of our
integrated provenance approach from Qd and Qo alluvial
deposits and modern drainages suggests Qo deposits are most
compatible with a Mission Creek source (or possibly a Morongo
Valley source for oldest strata) in the SBM. Qd deposits are
incompatible with a SBM source and more closely match the
eastern LSBM, where catchments expose mostly Cretaceous
and Jurassic plutons of the Sierra Nevada batholith. These data
provide evidence of continued dextral fault displacement along
the Mission Creek fault strand in the San Gorgonio Pass.
These findings are based on our provenance data which show
that modern drainages along the SBM and LSBM are dominated
by plutonic and gneissic lithic grains and clasts. Detrital zircon U-
Pb ages measured by LA-ICP-MS reveal Mesozoic (65-120 Ma
and 130-170 Ma), Mesoproterozoic (1.3-1.5 Ga), and
Paleoproterozoic (1.6-1.8 Ga) populations. Sandstone
compositions of the Qd are characterized by a transition from
basement uplift to dissected magmatic arc provenance fields and
increase in plutonic character over time. Detrital U-Pb data reflect
dominantly Mesozoic ages, with lesser 1.3-1.5 Ga, and 1.6-1.8
Ga peaks. Clasts compositions are mostly monzonite,
granodiorite, biotite gneiss, and hornblende diorite. Matrix
sandstone of the Qo is similar to underlying strata, but contains
progressively more metasedimentary (phyllite, quartzite, and
marble) and volcanic lithic grains and clasts over time. Higher
proportions of Proterozoic zircons and sparse Paleozoic zircons,
together with conspicuous quartzite and phyllite, may reflect
contributions from the Big Bear Group in the northern SBM. The
Jurassic age peak covaries with the abundance of hornblende
diorite clasts and plutonic lithic fragments, suggesting a
diagnostic signal of the diorite exposed in the western LSBM. We
note that the modern Whitewater River lacks diorite clasts,
precluding it as a source for the Quaternary deposits, all of which
contain diorite.
2016 SCEC Undergraduate Studies in Earthquake
Information Technology (UseIT): Earthquake Forecasting
Through Physics-Based Simulations, Zhenyu Fu, Morgan T
Bent, Hernan M Lopez, Spencer P Ortega, and Kevin C
Scroggins (Poster Presentation 329)
As part of the 2016 Undergraduate Studies in Earthquake
Information Technology (UseIT) internship program, students
worked in collaborative groups to tackle unsolved problems in
earthquake information technology presented in the form of a
Grand Challenge. Earthquake forecasting was the overall theme
of this year’s Grand Challenge. The UseIT High Performance
Computing (HPC) Team was challenged to simulate long
catalogs of California’s seismic activity on a supercomputer. The
team used a physics-based earthquake simulator, the Rate-State
earthQuake Simulator (RSQSim). The students ran the
simulations on the Blue Waters system at the University of
Illinois, one of the most powerful open-science supercomputers
in the world. To configure an RSQSim simulation, a series of
initial physical parameters must be specified, including initial
normal and shear stresses, rate- and state-friction parameters,
and earthquake slip rate. The HPC Team studied the effects of
these parameters on simulated catalogs. Each team member ran
several short simulations (25,000 simulated years) with different
input parameters and then compared the catalogs with the
Uniform California Earthquake Rupture Forecast Version 3
(UCERF3) to determine which parameter set produced the best
match. The HPC team also utilized the R Programing Language,
a language commonly used by statisticians as well as data
miners, in order to analyze the results of each catalog. Varying
the parameter sets produced drastic changes in the catalogs
generated. With the help of the Probabilistic Forecasting Team,
the HPC Team compared each of the short catalogs with multiple
aspects of the UCERF3 data. Of the seventeen short catalogs
generated, Sigma High had less than a 2% difference from
UCERF3 in the recurrence interval of events M ≥ 7 on the
Southern San Andreas Fault, which indicated that the parameter
set was the best representation of an earth-like system. Based
on this information, the team extended the catalog to 530,000
years in order to create a more comprehensive dataspace for
earthquake forecasting. The UseIT interns used these catalogs
to answer probabilistic questions posted in the Grand Challenge
and generate simulator-based forecasts for the San Andreas
Fault System.
Structure of the San Andreas Fault Zone in the Salton
Trough Region of Southern California: A Comparison with
San Andreas Fault Structure in the Loma Prieta Area of
Central California, Gary S Fuis, Rufus D Catchings, Daniel S
Scheirer, Mark R Goldman, and Edward Zhang (Poster
Presentation 005)
The San Andreas fault (SAF) in the northern Salton Trough, or
Coachella Valley, in southern California, appears non-vertical
and non-planar. In cross section, it consists of a steeply dipping
segment (75 deg dip NE) from the surface to 6- to 9-km depth,
and a moderately dipping segment below 6- to 9-km depth (50-
55 deg dip NE). It also appears to branch upward into a flower-
like structure beginning below about 10-km depth. Images of the
SAF zone in the Coachella Valley have been obtained from
analysis of steep reflections, earthquakes, modeling of potential-
field data, and P-wave tomography.
Review of seismological and geodetic research on the 1989 M
6.9 Loma Prieta earthquake, in central California (e.g., U.S.
Geological Survey Professional Paper 1550), shows several
features of SAF zone structure similar to those seen in the
northern Salton Trough. Aftershocks in the Loma Prieta
epicentral area form two chief clusters, a tabular zone extending
from 18- to 9-km depth and a complex cluster above 5-km depth.
The deeper cluster has been interpreted to surround the chief
rupture plane, which dips 65-70 deg SW. When double-
difference earthquake locations are plotted, the shallower cluster
contains tabular subclusters that appear to connect the main
rupture with the surface traces of the Sargent and Berrocal faults.
In addition, a diffuse cluster may surround a steep to vertical fault
connecting the main rupture to the surface trace of the SAF.
These interpreted fault connections from the main rupture to
surface fault traces appear to define a flower-like structure, not
unlike that seen above the moderately dipping segment of the
SAF in the Coachella Valley. But importantly, the SAF,
interpreted here to include the main rupture plane, appears
WELCOME
2016 SCEC Annual Meeting page 177
segmented, as in the Coachella Valley, with a moderately dipping
segment below 9-km depth and a steep to vertical segment
above that depth. We hope to clarify fault-zone structure in the
Loma Prieta area by reanalyzing active-source data collected
after the earthquake for steep reflections.
Preliminary late Pleistocene slip rate for the western Pinto
Mountain fault, Morongo Valley, southern California,
Katherine Gabriel, Doug Yule, and Richard V Heermance (Poster
Presentation 119)
The northern Coachella Valley region of the San Andreas fault
(SAF) system in southern California is complicated by
overlapping, active strands and its intersection with prominent,
secondary structures such as the Pinto Mountain fault. Recent
work in this area proposes that strain may be transferred from
the Mission Creek strand of the SAF to the Eastern California
Shear Zone (ECSZ), at least partly via the Pinto Mountain fault.
Like the better known Garlock fault, the Pinto Mountain fault is a
major east-west trending left-lateral transverse fault. It intersects
the Mission Creek strand of the SAF in the eastern San Gorgonio
Pass area, and its trace east through the Morongo Valley may be
a surface manifestation of strain transfer to the ECSZ.
Geodetically modeled slip rates for the Pinto Mountain fault vary
widely from 1 to 12.5 mm/yr, and geologic rates have been
speculative because the ages of possible offsets are
unconstrained. We have determined a preliminary geologic slip
rate from a left-lateral offset within alluvium in Big Morongo
Canyon in Morongo Valley. We measured a 205 +/- 20 m offset
of a strath contact between old alluvial gravels (Qoa) and
underlying gneiss (ggm). This set of piercing points assumes
pure left-lateral strike-slip motion along the main trace of the
western Pinto Mountain fault. A secondary fault trace may cut the
young alluvial gravels (Qa) to the north of the main trace, but any
lateral offset is unconstrained and therefore not included in our
rate calculation.
We obtained cosmogenic 10Be exposure ages of six monzo-
granite boulders on the surface of Qoa. Assuming zero-erosion
rate, boulder ages range from 64 ka to 90 ka with a preferred
average of 83 ka. Combining the offset and age measurements,
we calculate slip rates of 2.3 mm/yr to 3.2 mm/yr, with a preferred
slip rate of 2.5 mm/yr for the Pinto Mountain fault over the last
~90 ka. Although the Pinto Mountain fault is antithetic to the SAF,
we think that it likely represents a direct connection from the
Mission Creek strand of the SAF to the ECSZ. Furthermore, we
think that Pinto Mountain fault motion contributes to a decrease
in the slip rate on the SAF system as is passes northwest through
the San Gorgonio Pass region by transferring strain directly to
the ECSZ.
Asperity break after 12 years: The Mw6.4 2015 Lefkada
(Greece) earthquake, Frantisek Gallovic, Efthimios Sokos, Jiri
Zahradnik, Anna Serpetsidaki, Vladislav Plicka, and Anastasia
Kiratzi (Poster Presentation 055)
The Mw6.4 earthquake sequence of 2015 in western Greece is
analyzed using seismic data. Multiple point source modeling,
nonlinear slip patch, and linear slip inversions reveal a coherent
rupture image with directivity toward the southwest and several
moment release episodes, reflected in the complex aftershock
distribution. The key feature is that the 2015 earthquake ruptured
a strong asperity, which was left unbroken in between two large
subevents of the Mw6.2 Lefkada doublet in 2003. This finding
and the well-analyzed Cephalonia earthquake sequence of 2014
provide strong evidence of segmentation of the major dextral
Cephalonia-Lefkada Transform Fault (CTF), being related to
extensional duplex transform zones. We propose that the
duplexes extend farther to the north and that the CTF runs
parallel to the western coast of Lefkada and Cephalonia Islands,
considerably closer to the inhabited islands than previously
thought. Generally, this study demonstrates faulting complexity
across short time scales (earthquake doublets) and long time
scales (seismic gaps).
Monitoring of Microseismicity in the Peach Tree Valley
Region with Array Techniques, Jose Luis L Garcia-Reyes,
and Robert W Clayton (Poster Presentation 190)
This study is focused on the analysis of microseismicity along the
San Andreas Fault in the PeachTree Valley region with the use
of array techniques. This zone is located in the transition zone
between the locked portion to the south (Parkfield, CA) and the
creeping section to the north (Jovilet, et al., JGR, 2014). The goal
of the study is to relate the style of seismicity and its spatial
distribution to the mechanical state of the Fault. We use a 3D
backprojection technique and explore the use of Hidden Markov
Models to identify different patterns of seismic activity (Hammer
et al., GJI, 2013). The results show the evolution of
microseismicity as well as at least two different patterns of
seismic signals. The data for the study comes from a 2-week
deployment on July, 2014 of 116 single component nodes in a
cross-shaped configuration (8.2 km along and 9 km across the
Fault).
Aftershock Productivity on Volcanoes: What can it tell us
about interpreting aftershocks? Ricardo Garza-Giron, Emily
E Brodsky, and Stephanie G Prejean (Poster Presentation 255)
Most earthquakes have aftershocks and in some systems, it is
thought that the majority of the earthquake rate is comprised of
aftershocks. However, volcanic earthquakes are intrinsically
different from tectonic ones in that magmatic movement can
stress rocks more quickly than plate tectonics and elastic strain
can be more difficult to sustain over long periods of time. In this
study, we measure the aftershock productivity of 27 volcanoes in
the Alaska Peninsula and the Aleutian Islands. We utilize a
standard model of aftershock production developed for tectonic
systems where it is usually observed that the number of
aftershocks of a mainshock magnitude M is equal to K10 a(M-–
Mref) where K and a are constants that vary regionally and Mref
is a reference magnitude. Of the volcanoes studied here, 22
showed a mainshock-–aftershock behavior such as the one
observed in global or regional tectonic regimes. The remaining
volcanoes did not have enough seismicity for us to make a clear
assessment. Of those, 13 volcanoes are well–fit by the model
with p–values below 0.01. Those that are well–fit have a level of
aftershock productivity similar to an ordinary tectonic system like
Southern California (K=0.02). Intriguingly, the ones that are not
well–fit are open systems, while those that are well–fit are
primarily closed. To some extent, this distinction is a result of the
low seismicity in open systems, however, Shishaldin volcano has
high seismicity, poor aftershock clustering and is an open
system. It is a single case, but perhaps it suggests a more
general behavior. It is possible that only closed systems are able
WELCOME
2016 SCEC Annual Meeting page 178
to store enough elastic strain to make efficient aftershock
production. The volcanic results hint that aftershock productivity
may be a useful indicator of stress accumulation in the crust.
Blurring the boundary between earthquake forecasting
and seismic hazard, Matthew C Gerstenberger, David A
Rhoades, Graeme H McVerry, David S Harte, and Annemarie
Christophersen (Oral Presentation 9/13/16 14:30)
Earthquake forecasting and seismic hazard have been
traditionally considered as independent and separate fields of
study. We are currently working on a range of topics that are
beginning to blur the bounds between the two fields. With the
Canterbury earthquakes it became apparent that traditional
methods of seismic hazard modelling would not fully represent
the expected hazard in the coming 50 years and we developed a
hybrid long-term and time-dependent hazard model that melded
the two fields. A distinction of this model is that it combined a
range of models that produced forecasts from one-year to 50-
years. A major contribution to the estimated uncertainty in the
hazard came from the expected occurrence rate for the region in
20 to 50 years time. We are now developing models to allow
improved estimates of this rate and to better capture the
epistemic uncertainty in the short to long-term processes. By
using the hybrid model idea, we combine, in an alternative to
logic-trees, models based on differing data sets or hypotheses to
produce a forecast of earthquake occurrence. By combining
information from such things as seismicity, geodetic strain,
geological data and slow slip events, we can provide forecasts
that are typically more informative than any single model.
Additionally, we are trying to better characterize the uncertainty
inherent in all of the models; through an improved understanding
of this uncertainty, we can develop better statistical tests of the
models, and ideally provide more useful information to decision
makers who are using the outputs of such models both in the
form of short-term earthquake forecasts and long-term hazard
forecasts. Finally, through our experiences in Christchurch and
more recent earthquakes in New Zealand, we have gained
valuable experience in how to communicate earthquake forecast
information and, also, perhaps, a unique perspective on where
future earthquake forecasting and hazard research might be the
most beneficial to end-users of such information.
Strong along-strike variation in aftershock distribution and
rupture propagation of 2015 Mw 7.8 Gorkha earthquake
in Nepal, Abhijit Ghosh, Bo LI, and Manuel M Mendoza (Poster
Presentation 251)
Gorkha earthquake presents a unique opportunity to study active
tectonics, earthquake dynamics and fault structure in the
Himalayas in Nepal. We use four large aperture seismic arrays
at teleseismic distances to image the rupture propagation of the
mainshock and aftershock activities immediately following it. In
addition, a dense local network of 45 seismic stations is deployed
after the mainshock covering the entire rupture area to capture
the aftershock activities.
We calibrate and combine four large aperture seismic arrays to
image rupture propagation in unprecedented resolution. It shows
rupture complexities and branching near the end of this
damaging quake. Moreover, we solely use the arrays to detect
and locate aftershock activities immediately following the
mainshock. The aftershock distribution determined only by the
arrays matches well with the standard global earthquake catalog,
but detects more than twice as many aftershocks during the
same time period. It shows that existing seismic arrays alone can
be used to rapidly detect more aftershocks and determine their
spatiotemporal distribution to help hazard assessment and fast
response right after a destructive earthquake.
Furthermore, we use the dense local seismic network to detect
and locate aftershock activities with local magnitude as low as
0.3. Relative relocations of microseismicity provide a high-
resolution image of the fault structure in this area. They illuminate
a gently dipping decollement with several steeply dipping faults,
which are responsible for majority of the aftershock production.
Moreover, there is a sharp along-strike variation in aftershock
distribution, possibly related to the mainshock rupture and
underlying structure. This study is providing new insights into the
dynamics of the Gorkha earthquake, structures in the Himalayan
fault system and their implications in the seismic hazards in this
region.
Simulation Based Earthquake Forecasting with RSQSim,
Jacquelyn J Gilchrist, Thomas H Jordan, James H Dieterich, and
Keith B Richards-Dinger (Poster Presentation 316)
We are developing a physics-based forecasting model for
earthquake ruptures in California. We employ the 3D boundary
element code RSQSim to generate synthetic catalogs with
millions of events that span up to a million years. The simulations
incorporate rate-state fault constitutive properties in complex,
fully interacting fault systems. The Unified California Earthquake
Rupture Forecast Version 3 (UCERF3) model and data sets are
used for calibration of the catalogs and specification of fault
geometry. Fault slip rates match the UCERF3 geologic slip rates
and catalogs are tuned such that earthquake recurrence
matches the UCERF3 model. Utilizing the Blue Waters
Supercomputer, we produce a suite of million-year catalogs to
investigate the epistemic uncertainty in the physical parameters
used in the simulations. In particular, values of the rate- and
state-friction parameters a and b, the initial shear and normal
stress, as well as the earthquake slip speed, are varied over
several simulations. In addition to testing multiple models with
homogeneous values of the physical parameters, the parameters
a, b, and the normal stress are varied with depth as well as in
heterogeneous patterns across the faults. Cross validation of
UCERF3 and RSQSim is performed within the SCEC
Collaboratory for Interseismic Simulation and Modeling (CISM)
to determine the affect of the uncertainties in physical parameters
observed in the field and measured in the lab, on the
uncertainties in probabilistic forecasting. We are particularly
interested in the short-term hazards of multi-event sequences
due to complex faulting and multi-fault ruptures.
SCEC Web in the Clouds: Motivations and Experiences,
David J Gill, Philip J Maechling, Tran T Huynh, John E Marquis,
John Yu, Edric Pauk, Jason Ballmann, Mark L Benthien, Deborah
Gormley, and Annie Lo (Poster Presentation 328)
Last year, the SCEC.org website underwent a major overhaul in
terms of technology and appearance. We went from a custom
PHP solution to an open-source solution based on Drupal 7.
Basing the site on Drupal meant that we could use an actively
maintained open-source framework for the site. In addition, we
WELCOME
2016 SCEC Annual Meeting page 179
could use all the community-developed open-source modules to
handle everything from LDAP connections to file hosting on
Amazon Web Services. When the new Drupal 7-based SCEC
site was made live in September of 2015, we hosted the site on
local servers at USC. However, we recognized a variety of
challenges associated with hosting such a critical component of
our business infrastructure locally, including earthquake
resilience, backup infrastructure, power outages, and others
scenarios, that could result in the SCEC web site being down.
To help us address these problems, we have begun migrating
SCEC web infrastructure to the cloud. Unlike our previous self-
hosted solution at USC, cloud hosting has cost and scalability
considerations. We selected Amazon for our host as it is cost-
effective, has a tremendous infrastructure for scalability, and has
a robust suite of tools that enable us to create quality solutions
for the SCEC web infrastructure.
To support SCEC broad and growing web development needs,
we have created a web development environment that enables
us to deploy the current scec.org web site (with or without
content) to a local development machine, to a staging server, and
to the live site. Cloud hosting enables us to trivially ensure that
our staging server and live site have the exact same software
stack. Because they are using the exact same version of Linux
and the same backend infrastructure (S3 and RDS), we can
ensure that there are no errors that arise from different PHP
versions, different kernel versions, etc. when we transition new
features from the staging server to the live server.
Cloud hosting also allows for better compartmentalization. For
example, the database server need not be on the same server
as the Apache host. This provides for incredible scalability. We
used EC2 for the computing platform, RDS for the database, and
S3 for the file hosting. This allows for backups to be made with
regularity. Both our EC2 instance (which hosts all the code) and
RDS instance are backed up nightly. S3 has a feature that tracks
all changes made to a file. So if we accidentally overwrite a
document or an image, we can revert back to an earlier version.
This ensures that, for the most part, no matter what happens we
can always trivially revert the site back to a last known good state.
Constraining Interacting Fault Models in the Salton
Trough with Remote Sensing Data, Margaret T Glasscoe,
Jay W Parker, Andrea Donnellan, Gregory A Lyzenga, and Chris
W Milliner (Poster Presentation 152)
Constraining the distribution of slip and determining the behavior
of fault interactions is a complex problem. Field and remotely
sensed data often lack the necessary coverage to fully resolve
fault behavior. However, realistic physical models may be used
to more accurately characterize the complex behavior of faults
constrained with observed data, such as GPS, InSAR, and SfM.
These results will improve the utility of using combined models
and data to understand fault interactions, estimate earthquake
potential and characterize plate boundary behavior.
We will investigate how faults in the Salton Trough have been
affected by past several earthquakes by constructing viscoelastic
models and comparing them to observed geodetic data. The fault
segments and slip distributions are modeled using the JPL
GeoFEST software. GeoFEST (Geophysical Finite Element
Simulation Tool) is a two- and three-dimensional finite element
software package for modeling solid stress and strain in
geophysical and other continuum domain applications [Lyzenga,
et al., 2000; Glasscoe, et al., 2004; Parker, et al., 2008, 2010].
New methods to advance geohazards research using computer
simulations and remotely sensed observations for model
validation are required to understand fault slip, the complex
nature of fault interaction and plate boundary deformation. These
models help enhance our understanding of the underlying
processes, such as transient deformation and fault creep, and
can aid in developing observation strategies for sUAV, airborne,
and upcoming satellite missions seeking to determine how faults
behave and interact and assess their associated hazard.
The 2016 Mw5.1 Fareview, Oklahoma earthquakes:
Evidence for long-range poroelastic triggering at >30 km
from disposal wells, Thomas H Goebel, Matthew Weingarten,
Xiaowei Chen, Jackson Haffener, and Emily E Brodsky (Poster
Presentation 202)
Much of the surge in seismic activity in the central United States
has been linked to injection induced pressure changes on pre-
stressed faults. This type of pressure-induced seismicity requires
a direct hydraulic connection between injection wells and faults,
which is complicated when earthquakes are located several
kilometers beneath the injection zone and 10s of kilometers away
from the injection well.
Here, we employ numerical and analytical models to resolve
triggering mechanisms within the greater Fairview region,
Oklahoma. The study region shows strong evidence for injection-
induced seismicity, including an obvious lack of seismicity before
2013, followed by a rapid increase in background rates between
2014 and 2016. Most of the injection activity was concentrated
toward the northwest of the study region resulting in a relatively
cohesive zone of high-pressure perturbations between 0.1 and 1
MPa. This zone has a diameter of 10 to 20 km and produced
much seismicity most likely by direct pressure effects and fault
assisted pressure migration to larger depth.
Outside of the high-pressure zone, we observed two remarkably
detached, linear seismicity clusters at 30 to 60 km distance. Our
semi-analytical models reveal that poro-elastically-induced
Coulomb stress changes surpass pressure changes at these
distances, providing a plausible triggering mechanism beyond
direct pressure effects. Our results indicate that both pressure
and poroelastic stress changes can play a significant role in
triggering deep and distant earthquakes by fluid injection and
should be considered for seismic hazard assessment beyond the
targeted reservoir.
Published Quaternary geochronologic data show the
importance of dating geomorphic surfaces with multiple
geochronometers, Peter O Gold, and Whitney M Behr (Poster
Presentation 090)
Quaternary fault slip rates rely on geochronologic dating methods
for interpreting the ages of displaced geomorphic surfaces such
as alluvial fans and fluvial terraces. Most commonly employed
among these are cosmogenic exposure dating, uranium-series
dating, optically stimulated luminescence, and radiocarbon
dating. Because the geologic samples collected may have
experienced very different pre-, syn-, and post-formation
WELCOME
2016 SCEC Annual Meeting page 180
histories, the dates may be interpreted as max ages, min ages,
or something in between, with implications for whether a fault slip
rate should be considered an upper or lower limit. For this
reason, the practice of using multiple geochronometers to
constrain a range of possible surface ages and slip rates is ideal
and has become common enough to begin assessing how well
dates obtained using different methods agree. To this end, as
well as to assess the typical scatter within populations of
exposure dates, we have compiled data from >120 published
studies in all reporting >1500 dates of different types from >300
alluvial fans and fluvial terraces. Where two dating methods were
used, it was most common to compare exposure dates to those
obtained using another method. In such cases, the exposure
dates were almost without exception older (likely due to inherited
nuclides), and for only ~50% of these studies did the dates differ
by an amount attributable to different depositional processes. On
average, the standard deviation of exposure dates from clasts is
30-40% of the median age, indicating a degree of scatter that is
typically addressed by excluding, on average, 25% of dates as
outliers. Inheritance in exposure ages from clasts remains
challenging estimate, and only in rare cases does subtracting the
nuclide concentration found in the modern wash from that
measured on surface bring the latter into concordance with
independent dates. Cosmogenic depth profiles are useful for
estimating inheritance in the clast size sampled, but 2 in 3 depth
profiles exhibited distributions too scattered for a date to be
calculated. Given uncertainties related to depositional process
and internal scatter in individual datasets, a review of published
Quaternary geochronologic data shows the importance of dating
geomorphic surfaces with more than one method in order to
derive a more realistic range of surface ages.
Seismogeodetic Observations of the June 10, 2016 M5.2
Borrego Springs Earthquake, Dara E Goldberg, Jessie K
Saunders, Jennifer S Haase, and Yehuda Bock (Poster
Presentation 256)
Seismogeodesy is the optimal combination of high rate GPS
observations and strong motion accelerations in a Kalman filter
to produce broadband velocity and displacement waveforms,
including accurate measure of the static offset. This approach
requires collocated GPS and accelerometer instrumentation. Our
group developed a low cost geodetic module and MEMS
accelerometer package (SIO GAP) that is installed on the vertical
leg of existing cGPS stations and whose data is communicated
along with the GPS data in real time. So far, SIO GAPs have
been installed at 17 PBO/SCIGN stations spanning the San
Andreas, San Jacinto, and Elsinore Faults and at 10 PBO
stations in the San Francisco Bay Area.
On June 10, 2016 a M5.2 earthquake occurred in Borrego
Springs, California, near the San Jacinto Fault. The earthquake
was widely felt in southern California, and was observed on
seismogeodetic instrumentation located 17 km to 100 km from
the source. We present a retrospective analysis of the data and
early warning products available from the collocated sites.
We compare seismogeodetic combinations using observatory-
grade accelerometers to those using the SIO GAPs to
demonstrate the capabilities of these low-cost instruments in the
field. While the collocated Piñon Flat station PIN2 does not
measure the anomalously high PGA observed at nearby
observatory-grade strong motion station PFO, the
seismogeodetic waveforms exhibit similar characteristics, with
particularly good agreement in the velocities. We show the utility
of the seismogeodetic velocities for detection of first seismic
arrivals as well as the results of our rapid event location and
shaking onset prediction as part of our prototype seismogeodetic
early warning system. From the location estimate, we estimate
the moment magnitude using a simple scaling relation from peak
ground displacement measured from the seismogeodetic
displacement.
This event was also observed on seismogeodetic
instrumentation used for structural monitoring at a parking
garage at the Automatous University of Baja California Faculty of
Medicine in Mexicali. The observations show the building
response, and demonstrate the feasibility of structural monitoring
using seismogeodesy. No discernable displacement was
measured at 126 km epicentral distance, however the MEMS
accelerometers were available to provide an estimate of building
response with rooftop amplification of 1.5 times in the east
component and 2.3 times in the north component.
Seismic Evidence for Splays of the Eureka Peak Fault
beneath Yucca Valley, California, Mark R Goldman, Rufus D
Catchings, Joanne Chan, Robert R Sickler, Coyn Criley, Dave
O'Leary, and Alan Christensen (Poster Presentation 262)
In April 2015, we acquired high-resolution P- and S-wave seismic
data along a 3.1-km-long, E-W-trending profile in Yucca Valley,
California. Our seismic survey was designed to locate possible
sub-parallel faults of the Eureka Peak Fault, which trends ~NW-
SE near the western end of our profile. The Eureka Peak Fault is
a potential hazard to the Yucca Valley region, as it appears to
have experienced surface ruptures associated with both the 23
April 1992 M 6.1 Joshua Tree earthquake and the 28 June 1992
M 7.3 Landers earthquake. We simultaneously acquired P- and
S-wave data using explosive sources spaced every 100 m, along
with higher resolution P-wave data from seisgun sources spaced
every 5 m. Each shot was co-located with and recorded by 634
P-wave geophones (40-Hz) spaced 5 m apart and 250 S-wave
geophones (4.5-Hz) spaced 10 meters apart. We developed P-
wave tomographic velocity models and reflection images that
show at least one significant fault about 2.3 km NE of the Eureka
Peak Fault. This fault may potentially pose a hazard and affect
groundwater flow in the area.
Guadalupe Island as a constrain for earthquake hazard at
Californias's shoreline, Jose Javier Gonzàlez-Garcìa,
Alejandro Gonzalez-Ortega, John E Galetzka, Christian Walls,
Hebert Martinez-Barcena, and Juan C Robles (Poster
Presentation 146)
For the purpose of tectonic-earthquake hazard in the Pacific-
NorthAmerica western border region, we look for high accuracy
of relative plate boundary instead of having global coverage (e.g.
GEODVEL, MORVEL56, ITRF2008PPM, GSRM2); by imposing
Guadalupe Island GPS velocity to be part of the Pacific Plate
motion model. We obtain updated GPS velocity at Guadalupe
Island in 3 surveyed campaign sites: GAIR, GUAD and RMGU,
and the permanent site GUAX recently upgraded in June 2016.
We process the data obtained since 1999 to date using
GAMIT/GLOBK and align the results to ITRF2008 using IGS. Our
WELCOME
2016 SCEC Annual Meeting page 181
preliminary velocity (mm/yr) for GUAX is 24.9n, -47.2e and -0.21
up.
Mean horizontal velocity for Guadalupe is 25.0n, -47.1e,
consistent with our former results and with
GUAX.pbo.nam08.pos. Our model for Pacific Plate motion,
including our 4 GPS sites at Guadalupe and 23 published
velocities (DeMets, et al, 2014) from other Pacific Ocean islands
is -62.63n,107.89e and 0.677°/Myr. Residuals (mm/yr) for
emblematic sites in the California shoreline are: FARB -
1.2n,0.4e; HARV -1.1n,-0.3e; P172 -1.6n,-0.2e and VNDP -2.2,-
0.5. Then, there are minor earthquake hazard for offshore faults
west of this sites.
Comparison of our GPS velocity at Guadalupe Island with recent
global models shows no residual for GSRM2, and only -0.9n,
0.6e for NNRMORVEL56. This indicate that Guadalupe Island
belongs to the Pacific Plate for as both geodetic and geologic
models between 1mm/yr bounding error.
Investigating tremor sources along the San Andreas Fault
using integrated static and dynamic stress models, Hector
Gonzalez-Huizar, Sandra Hardy, and Bridget Smith-Konter
(Poster Presentation 179)
Ambient and triggered tectonic tremors have been reported near
the creeping to locking transition zone along the Parkfield-
Cholame section of the San Andreas Fault, as well as in the San
Jacinto and Calaveras Faults. Tremors dynamically triggered by
seismic waves can be useful in the estimation of the friction and
stress conditions at aseismic depths. These are important
parameters in computing stress transfer from plate motion to the
seismogenetic zones, and thus, in creating seismic hazard
models. We use recorded seismic signal from large earthquakes
to calculate the dynamic stresses capable to trigger tremor in
these regions. Then, dynamics stresses are integrated with local
tectonic stress models with the objective to estimate the spatial
variability of frictional and stress parameters along the areas
where tremors are triggered. Integrating static and dynamic
stress along the San Andreas Fault allows us to better
understand the physical conditions necessary for tremor
occurrence.
Geodetic moment accumulation and seismic release in
northern Baja California, Mexico, Alejandro Gonzalez-
Ortega, Jose Javier Gonzàlez-Garcìa, and David T Sandwell
(Poster Presentation 155)
We have analyzed all available GPS data from northern Baja
California, Mexico, covering the period 1993-2008, in order to
estimate the strain rate and seismic moment accumulation rate
during the interseismic period prior to the 2010, Mw 7.2 El Major-
Cucapah earthquake. The analysis shows concentrations of
strain-rate and moment-rate along the Imperial and Cerro Prieto
Faults extending from the Salton Sea to the Gulf of California.
The regional moment accumulation based on geodesy (M _̇o^g=
3.0±0.3x1018Nm/yr) is significantly greater than the
corresponding moment release rate by earthquakes
(M _̇o^s=1.2x1018 Nm/yr), so in 2008 the region had a residual
moment of Mo= 2.1x1020 Nm, equivalent to an Mw7.4
earthquake. Much of this excess moment was released by the
2010, Mw7.2 El Mayor-Cucapah earthquake; the remaining
residual moment of the Imperial and Cerro Prieto faults
corresponds to a maximum Mw6.8 according to the Gutenberg–
Richter law with b-value of 1.0. In contrast, the seismic moment
released on the San Miguel fault, primarily from the earthquake
swarm occurred in 1954-1956 Mw<=6.8, is one order of
magnitude higher than the accumulated geodetic moment.
These new estimates of seismic moment potential provide
important constraints on earthquake forecasts for the southern
San Andreas Fault System.
Laboratory Earthquakes Nucleated by Fluid Injection,
Marcello Gori, Vito Rubino, Ares J Rosakis, and Nadia Lapusta
(Poster Presentation 035)
Fluids play an important role in earthquake source processes.
Numerous field observations have highlighted the connection
between fluids and faulting in triggering events ranging from
earthquakes to creeping motion. Our newly developed laboratory
setup allows the rapid injection of fluid onto a frictional interface
in an acrylic specimen loaded in compression and shear. The
fluid injection has controlled fluid pressure, time, and flow rate.
We have successfully nucleated dynamic ruptures by fluid
injection and captured their propagation using our rich collection
of diagnostics encompassing: (1) high-speed digital image
correlation (DIC) for full field measurements of displacements,
velocities, strains and stresses; (2) strain gage array for high-
temporal resolution point-wise strain measurements; and (3)
laser velocimetry for high-temporal resolution point-wise velocity
measurements. Hence we can nucleate dynamic ruptures via the
injection of fluid in a controlled laboratory environment. We plan
to use the setup to explore the range of experimental parameters
that leads to triggering dynamic events and/or slow slip.
How we learned to stop worrying and start loving bulk
nonlinearities, Setare Hajarolasvadi, and Ahmed E Elbanna
(Poster Presentation 041)
The Finite Difference (FD) and Boundary Integral (BI) methods
have been extensively used in computational earthquake
dynamics. The FD method provides a powerful tool for simulating
a variety of rupture scenarios including nonplanar faults and bulk
plasticity. However, it requires the discretization of a significant
portion of the bulk to avoid artificial boundary reflections. This
makes simulation of large scale ruptures computationally
prohibitive. For the BI method, the explicit numerical
representation is confined to the crack path only, with the bulk
elastodynamic response expressed in terms of integral relations
between displacement discontinuities and tractions along the
crack. This leads to significant computational saving especially
when the spectral formulation of the method [Lapusta et al. 2000]
is used. The spectral boundary integral equation method [SBIE],
however, works only for linear elastic bulk and planar fault
surfaces. Here, we propose a novel hybrid numerical scheme
that combines FD and the SBIE to enable treating fault zone
nonlinearities and heterogeneities with higher resolution and in a
more computationally efficient way.
In the proposed method, we enclose the near-fault bulk
inhomogeneities (e.g. non-planar faults surface, off-fault plastic
zone, or low velocity zones) in a virtual strip that is introduced for
computational purposes only. Only this strip is then discretized
using the finite difference method while the virtual boundaries of
the strip are handled using the independent formulation of the
SBIE method which simulates each half space independently.
WELCOME
2016 SCEC Annual Meeting page 182
The finite difference solution of the strip provides traction to the
virtual boundaries. These tractions are used as input for the SBIE
to compute the displacements which are in turn applied to the
virtual strip as Dirichlet boundary conditions to advance the
solution to the next time step. We illustrate the accuracy and
efficiency of the method using several examples. By leveraging
the flexibility of bulk methods and the computational superiority
of boundary methods we expect the hybrid approach to enable
the integration of high resolution fault zone physics in
elastodynamic simulations and to facilitate cycle simulations of
faults in the presence of bulk nonlinearities at a reasonable
computational cost.
A derivation of the median ratios between different
definitions of horizontal component of ground motions in
Central and Eastern United States, Alireza Haji-Soltani, and
Shahram Pezeshk (Poster Presentation 263)
A single ground motion intensity measure, typically spectral
acceleration (SA), is required as the main input in deriving
empirical Ground Motion Prediction Equations (GMPEs). There
has been a debate for years on which of the two horizontal
ground motion orientations would be the most appropriate single
horizontal orientation to be used in developing GMPEs.
Traditionally, a single horizontal orientation has been used in
calculating SA for all periods. The spectrum changes with
orientation and using a single orientation to represent two
dimensional ground motions may lead to loss of this information.
Therefore, different techniques have been proposed in the
literature to combine the two horizontal components of ground
motions into a scaler definition. Several researchers have studied
the ratios between different definitions of horizontal component
of ground motions. While the most recent studies produced
similar results using different subsets of the NGA West-2
database, it is possible that such ratios may differ for different
regions. The purpose of this study is to derive ratios between
median values for different definitions of horizontal component of
ground motions in central and Eastern United States (CEUS)
using a subset of the NGA-East database. These median ratios
can be used along with the magnitude and distance scaling
factors in a coherent and consistent way when multiple GMPEs
are combined in a logic-tree framework in performing
Probabilistic Seismic Hazard Analysis (PSHA).
The Ups and Downs of Southern California: Mountain
Building, Sea Level Rise, and Earthquake Potential from
Geodetic Imaging of Vertical Crustal Motion, William C
Hammond, Geoffrey Blewitt, Corné W Kreemer, Reed J Burgette,
Kaj M Johnson, Charles M Meertens, and Frances Boler (Oral
Presentation 9/13/16 10:30)
Contemporary tectonic uplift in California and Nevada is an active
part of ongoing plate boundary processes driving earthquakes.
However, it has so far been difficult to confidently resolve and
interpret uplift patterns. The challenges are twofold. First is the
geodetic problem of isolating the signal of crustal-scale vertical
motion given a large number of noisy time series from multiple
networks irregularly distributed in space and time. Second is the
problem of partitioning the signals into patterns of long-term
tectonic deformation, earthquake cycle, flexural/isostatic
adjustment, local basin response from groundwater hydrology,
crustal loading, and/or mantle flow.
The spread of precise GPS networks and new developments in
processing GPS data are leading to a clearer and more complete
picture of vertical motions, improving knowledge of the rates of
mountain growth, associated fault slip, and seismic hazard. To
address the explosion in the quantity of new measurements, the
‘Plug and Play’ initiative is easing access to these data for
everyone, enhancing their utility and impact. Plug and Play
removes barriers at the beginning and end of the GPS
processing chain, providing a free GPS data processing service
and back-end products with free and open access. Currently the
system has global scope, providing products for ~15,000
stations. To derive uplift maps from the data we apply our new
robust estimation method "GPS Imaging" that combines non-
parametric robust trend estimation with median-based
despeckling and spatial filtering to suppress noise.
The resulting images show that the Sierra Nevada is the most
rapid and extensive uplift feature in the western United States,
rising up to 2 mm/yr, with uplift strongly modulated by climatic
and anthropogenic forcing from groundwater pumping. The
images reveal a discontinuity in the uplift field across Owens’
Valley, suggesting that Sierra Nevada uplift is associated with
crustal extension in the southern Walker Lane. Across the
Western Transverse Ranges (WTR) of southern California we
have taken the analysis further by combining four geodetic
techniques: GPS, InSAR, levelling and tide gauges to constrain
the vertical rate field. These results reveal 1-2 mm/yr of uplift
across the WTR and San Gabriel Mountains block, focused west
of the San Andreas fault, consistent with upward interseismic
extrusion of these blocks as they experience contraction against
the San Andreas Fault.
Late Quaternary slip rates from offset alluvial fan surfaces
along the Central Sierra Madre fault, southern California,
Austin Hanson, Reed J Burgette, Katherine M Scharer, and
Nikolas Midttun (Poster Presentation 096)
The Sierra Madre fault (SMF) is an east-west trending reverse
fault system along the southern flank of the San Gabriel
Mountains near Los Angeles, California. The ~140 km long SMF
is separated into four segments; we focus on the multi-stranded,
~85 km long Central Sierra Madre fault (CSMF). The CSMF lacks
a well-characterized geologic slip rate, and the goal of our work
is to constrain the slip rate of the CSMF using offset alluvial fan
terraces dated by terrestrial cosmogenic nuclide (TCN) and
infrared-stimulated luminescence (IRSL). Our main study area
includes three catchments along the western CSMF, where we
modified previous mapping (Crook et al., 1987) using field
observations and analysis of lidar DEMs to delineate terrace
flights in greater detail. Vertical separation estimates are derived
from 30-40 m wide swath profiles of topography extracted from
0.5-3 m resolution lidar DEMs, analyzed in 2 m wide subswaths.
Dip-slip displacements with uncertainties are calculated following
the methods of Thompson et al. (2002), which include a range of
estimates for the vertical separation, hanging wall and footwall
surface slopes, fault dip, and the location of the fault tip relative
to the scarp surface. For four surfaces we have TCN depth
profiles and complementary IRSL ages that provide an
independent maximum age of the deposit and/or a minimum age
from fine-grained material that post-dates terrace emplacement.
For three additional terraces, we have IRSL but no TCN ages.
Preliminary analysis of TCN data yields model ages that range
WELCOME
2016 SCEC Annual Meeting page 183
from ~20-65 ka and are consistent with the order of terrace
abandonment. IRSL indicate that the fine-grained material on top
of the terraces was deposited rapidly following terrace
abandonment. Preliminary slip rate estimates are ≤ 1 mm/yr in
two locations along the western CSMF. Although averaged over
a longer time period than previous studies, this slip rate may only
span 3-4 earthquake cycles suggested by paleoseismic work on
the CSMF. Our slip rate estimates are at the lower end of the
UCERF3 estimate of 1-3 mm/yr, which suggests that some
portion of the strain budget assigned to the CSMF has been
accommodated elsewhere over the late Quaternary.
Can tectonic loading be observed as interseismic stress
rotation? Jeanne L Hardebeck (Poster Presentation 014)
The shear stress on major faults evolves through the seismic
cycle, due to tectonic stress loading, coseismic stress release,
and earthquake stress transfer. If the seismic cycle stresses are
small compared to the background differential stress, the stress
orientations should not change during the seismic cycle.
However, observed coseismic stress rotations imply that the
stress drop is on the order of the differential stress. The
coseismic stress rotations suggest that the stress rotates back
during the rest of the seismic cycle as the fault is reloaded, raising
the possibility that monitoring interseismic stress changes could
inform earthquake hazard assessment. I test whether observable
interseismic stress rotations in southern California are consistent
with tectonic loading. I invert the focal mechanism catalog of
Yang et al (BSSA, 2012) for stress orientations in 4 time periods,
and look for significant changes in the direction of the maximum
horizontal stress axis, SHmax. For a simple loading model,
increased shear stress on strike-slip faults should correspond to
SHmax rotating towards a 45° angle to the fault strike. For the
San Andreas, San Jacinto, Elsinore, and Garlock faults,
however, >40% of sample points along the fault experience
SHmax rotating away from 45°. To better account for the
complexity of loading of the fault system, I compute the SHmax
rotation directions predicted by the SCEC Community Stress
Model (CSM). I add 33 years of loading from a stressing rate
model to a stress model, for different pairs of CSM models, and
compute the direction of SHmax rotation. Most pairs of models
exhibit similar patterns of SHmax rotation, featuring counter-
clockwise rotations centered along the major faults. The
observed rotations, in both directions, do not qualitatively match
these predicted patterns. I conclude that the interseismic tectonic
stress loading in southern California is not detectable, at least
over the 33-year time period of the mechanism catalog.
Aiming for Validation – The SCEC/USGS Dynamic
Earthquake Rupture Code Comparison Exercise, Ruth A
Harris (Poster Presentation 047)
The SCEC/USGS Dynamic Earthquake Rupture Code Group is
an international collaboration among scientists who use 3D
spontaneous rupture computer codes to numerically simulate
physics-based earthquake rupture and the resulting ground
motions. In the years to date, group members have tested the
viability of their computer codes by comparing the results
produced by each code, using an expansive set of benchmark
exercises. These exercises implement the range of assumptions
frequently used by modelers when they simulate earthquakes,
including heterogeneity in initial stress conditions, a variety of
formulations for fault friction, heterogeneous fault geometry
including 3D rough faults, branched faults, and fault-stepovers,
1D and 3D velocity structures, and off-fault yielding. Whereas
each exercise has sharpened the abilities of the codes and
enhanced our confidence in the results that they produce, we
have not yet rigorously tested the products of these codes
against observations from real earthquakes. In 2016 we have
started on this path toward code-validation, and our first test is
the 2000 Mw6.6 Tottori, Japan earthquake.
Characterizing Seismicity with High-Precision
Relocations of Recent Earthquake Sequences in Eastern
California and Western Nevada, Rachel L Hatch, Daniel T
Trugman, Ken D Smith, Peter M Shearer, and Rachel E
Abercrombie (Poster Presentation 225)
Recent earthquake sequences and swarms in eastern California
and western Nevada (2010 to present) have caught the attention
of the public and emergency responders. Our objective is to
characterize this seismicity by computing high-quality event
locations and source parameters, in order to resolve fault
structures and assess hazard implications. We obtain high-
precision relocations by applying the “GrowClust” algorithm
(Trugman et al., 2016; see presentation at this meeting) to
earthquake clusters at Herlong, CA (230+ events; 2016-present;
largest event = Mw 4.5) and Thomas Creek, S. Reno (183
events; 2015-2016; largest event = Mw 4.3). Relocations provide
evidence of complex structures within an overall transtensional
stress field in the Northern Walker Lane. The August 3, 2016 Mw
4.5 Herlong earthquake occurred at 9:55 pm local time and was
widely felt throughout the northeastern California and Reno-
Tahoe area. Relocations suggest the event was on a steeply NE
dipping structure striking ENE-WSW, possibly located on a
segment of the NW-SE striking Honey Lake fault of the Northern
Walker Lane. The regional surface-wave moment tensor solution
for the mainshock shows a well-constrained strike-slip
mechanism striking N50W (consistent with the orientation of the
Honey Lake fault) and dipping slightly NE at 83 degrees.
Additional short-period focal mechanisms also show evidence of
strike-slip faulting in this sequence. The Mw 4.3 December 22,
2015 Thomas Creek earthquake was also widely felt throughout
the Reno-Tahoe area. Relocations indicate a NNE-striking, west-
dipping normal faulting mechanism, implying the event occurred
along the Virginia Range frontal fault. This interpretation
suggests active normal faulting bounds both the eastern and
western Reno basin in the south Reno area in a graben structure.
The regional surface-wave moment tensor solution for the main
event shows a well-constrained normal-faulting mechanism
striking N17E and dipping 53 degrees west, consistent with the
dip of the aftershock distribution. Additional short-period focal
mechanisms computed using the HASH program (Hardebeck
and Shearer, 2002) show evidence for normal and oblique-
normal faulting in the foreshock and aftershock periods. From the
GrowClust relocation results, we estimate a rupture extent of 1.1
km during the Mw 4.3 mainshock, implying a stress drop of about
9 MPa. Our aim is to understand and characterize seismicity in
the Northern Walker Lane region. Relocation results for other
recent sequences generally show well-defined primary fault
planes with additional, complex off-fault structures that are well
resolved within the smaller sequences.
WELCOME
2016 SCEC Annual Meeting page 184
Progress towards deriving the three-dimensional
coseismic deformation field along the White Wolf fault
during the Mw ~7.3 1952 Kern County earthquake,
Alexandra E Hatem, James F Dolan, and Chris W Milliner (Poster
Presentation 017)
We present preliminary results of our efforts to constrain
coseismic surface deformation during the Mw~7.3 1952 Kern
County earthquake. We are using a modified workflow to
difference point clouds derived from legacy air photos taken
before and after the July 21 event, with the pre-event survey on
May 23, 1952 and the post-event survey spanning August to
September, 1952. Because no camera parameters were
recorded during the aerial photo surveys, we cannot use change
detection algorithms (e.g., COSI-Corr) that have previously been
used to measure coseismic surface deformation in other
earthquakes. Instead, we create point clouds from the
unreferenced and uncalibrated pre- and post-event images using
Agisoft Photoscan. We georeference the air photos using a
modern (Landsat) image, which enables the creation of a
georeferenced point cloud, as well as a digital surface model and
orthorectified, mosaicked image. Pre- and post-earthquake
georeferenced point clouds are then differenced using the
iterative closest point (ICP) algorithm within Cloud Compare. This
allows us to derive the x, y, and z cloud-to-cloud distances, and
hence the cosesimic deformation field. We have conducted a test
of the workflow on the eastern part of the Kern County rupture
along sections of the White Wolf fault where surface rupture was
documented by Bulwada and St. Armand (1955). We observe
primarily thrust displacement (south-side-up), and in some
instances we detect near-field warping on the order of 1-4
meters. Creating a complete, three-dimensional surface
deformation field along this rupture will provide a much-needed
data set of ground deformation during a large-magnitude
earthquake generated by rupture of a partially blind thrust fault
that exhibits multiple senses of displacement. The deformation
field associated with the 1952 White Wolf fault event will provide
a useful analog for similar deformation in future large-magnitude
earthquakes generated by the major blind thrust faults beneath
metropolitan Los Angeles, including the Puente Hills and
Compton blind thrust faults.
Optical Fiber Strainmeters: Developing Higher Dynamic
Range and Broader Bandwidth, Billy Hatfield, Mark A
Zumberge, Frank K Wyatt, and Duncan C Agnew (Poster
Presentation 261)
We have been developing near-surface horizontal strainmeters
and vertical borehole strainmeters that use passive optical fibers
as their primary sensing element: optical-fiber strainmeters, or
OFS. These instruments are useful in two areas. First, the
sensing of deformation over long baselines (hundreds of meters),
without calibration issues, over a band from 100 Hz to a few days:
a potential alternative to borehole strainmeters, especially given
lower cost and usefulness in a wide range of geologic settings.
Second, use in Earthquake Early Warning (EEW), where the
bandwidth provides a single record including both the dynamic
signals from radiated waves and the near-field static offset, both
of which are important in EEW; the long baselength allows more
reliable measurement of the offset. The optical-fiber systems
have a dynamic range that is
fundamentally limited only by the fiber becoming nonlinear;
between this and the resolution of the system the range is 120
dB or better. The current systems are limited by the ability of the
fringe-monitoring system to reliably track very high rates of strain;
this occurred on the fiber strainmeters installed at Pinon Flat
Observatory (PFO) at the time of the recent Anza earthquake in
June (2016:162, Mw 5.2, hypocentral distance 23 km), though
the systems behaved linearly otherwise. We are, with SCEC
funding, developing a fringe-counting system with very high rate
fringe tracking. For broadening the bandwidth at low frequencies,
temperature-induced noise must be reduced; we have developed
an OFS using a dual-fiber system, measuring changes in optical
path length along two fibers with different temperature
coefficients; the resulting time series can be combined to provide
a record of the strain much less affected by temperature. We will
report primarily on the performance of the OFS systems installed
at PFO, where we can compare them with the existing vacuum-
pipe strainmeters for microseisms, earthquakes, tides, and
longer periods. But we also have useful results from the first
deployment of a long-base strainmeter on the sea floor, at a
depth of 1900 m and 93 km west of the Oregon coastline.
Rapid Rupture Directivity Determination of Moderate Dip-
Slip Earthquakes with the Reduced Finite Source
Approximation, Xiaohui He, and Sidao Ni (Poster Presentation
211)
Moderate (M5.5-7) dip-slip earthquakes are frequent in both
interplate and intraplate regions, and some of them could cause
severe damage to buildings and even casualties. For an
earthquake with up-dip rupture, the hanging wall region usually
experiences enhanced strong ground motion significantly, and
may suffer more damage. Determining the fault plane is thus
essential for rapid hazard assessment, and rupture directivity
analysis based on seismic waveforms is an effective method to
resolve the ruptured plane as well as the rupture direction.
For strike-slip earthquakes which usually feature horizontal
rupture propagation, rupture directivity could be determined via
the spatial difference between hypocenter and centroid. But this
technique is usually not feasible for dip-slip earthquakes with
along dip rupture due to the large uncertainty of hypocenter depth
in routine location process. For dip-slip earthquakes, aftershock
distribution and finite fault inversion are commonly used to
resolve the rupture directivity. In regions with dense seismic
network, the spatial distribution of early aftershocks and the
relative focal depths for mainshock and aftershocks may indicate
the ruptured plane and the rupture direction, respectively.
However, the strike of the two nodal planes are approximately
same for dip-slip earthquakes, thus accurate focal depths for
aftershocks are required to distinguish the fault plane from the
auxiliary plane, which are difficult to obtain when the network is
sparse. Slip distribution and rupture process can be derived in
finite fault inversion, but the rapid application to moderate
earthquakes is limited due to the low resolution of teleseismic
waves or sparse local seismic network.
Instead, the teleseismic P wave can be split into the down-going
direct P-wave and up-going depth phases (pP and sP). When the
earthquake ruptures along dip, there is opposite rupture
directivity effects on direct P-wave and depth phases. For
instance, up-dip rupture will result in smaller amplitude and
WELCOME
2016 SCEC Annual Meeting page 185
longer duration for direct P-wave, larger amplitude and shorter
duration for pP and sP. Furthermore, it usually takes less than 15
minutes for teleseismic P wave to travel, therefore, teleseismic P
wave can be used in rapid rupture directivity analysis. We
proposed to calculate the reduced finite source (point source with
rupture directivity) synthetics for teleseismic P wave, and
determine the rupture directivity via waveform fitting. We verified
the effectiveness of this method with forward tests, and
investigated its robustness against station selection,
uncertainties in point source parameters and uncertainties in
finite source parameters. Then, we applied this method to the
2011 Mw5.8 Virginia earthquake and the 2008 Mw6.0 Nevada
earthquake, and obtained consistent results with previous
studies, suggesting our method is effective in rapid rupture
directivity analysis for moderate dip-slip earthquakes.
How Sensitive are Inferred Stresses and Stressing Rates
to Rheology? Clues from Southern California Deformation
Models, Elizabeth H Hearn (Oral Presentation 9/12/16 14:00)
Variations in fault zone strength and lower lithosphere
viscoelastic rheology have been shown to influence deformation
patterns in fault systems over a range of time scales (e.g. Bird
and Kong, 1994; Fay and Humphreys, 2005). Because of our
mandate to characterize stresses in the southern California
lithosphere, SCEC has proposed a resource (the Community
Rheology Model) that will facilitate representing Southern
California’s rheology in deformation models. I will present some
recent modeling results demonstrating how assumed fault zone
and rock rheologies influence estimates of stresses, stressing
rates and fault slip rates.
My kinematic and dynamic finite-element models incorporate the
UCERF3 block model-bounding fault geometry (Field et al.,
2014), with refinements suggested by geologists involved in the
UCERF3 effort. The kinematic models are fit to either the SCEC
CMM4 velocity field or a strain energy minimization criterion by
varying fault slip rates, within the UCERF3 ranges. The
(preliminary) dynamic models are tuned to fit geological fault slip
rates and SHmax orientations by varying fault zone and rock
rheologies. Deformation is driven from the sides in both sets of
models.
The kinematic models suggest that 22-35% of the strain
accumulation across Southern California is not associated with
interseismic locking of the modeled faults, consistent with other
studies (e.g. Bird, 2009, Johnson, 2013; Zheng and Shen, 2016).
This is true whether or not the GPS velocity field is corrected for
possible seismic cycle perturbations due to large SAF
earthquakes. Plasticity exerts a profound influence on slip rates,
deformation patterns and stresses obtained from dynamic
deformation models. Variation in fault zone strength (quantified
as shear traction) also affects stress magnitudes and
orientations, but not surface velocities or slip rates in models with
just the San Andreas Fault zone represented. Simulations are
underway to assess how variations in fault zone strength and
lithosphere rheology affect deformation in dynamic models
incorporating the full fault network.
NGA 2 GMPE’s Under Predict Long-Period Near-Source
Motions from Large Earthquakes, Thomas H Heaton, and
Becky Roh (Poster Presentation 276)
The recently constructed San Bernardino Law and Justice Center
is an 11-story story steel building located 6 km from the San
Andreas Fault. In order to mitigate strong shaking expected
during the lifetime of this expensive structure ($400M), the
designers chose to use base isolation to limit the inertial forces
from future earthquakes. The isolation system uses a triple
pendulum design with an approximate period of 5 ½ s and a
maximum isolator displacement of 1.1 m. The designers have
advertised that the building is designed for the maximum ground
motion expected with a repeat of 2,500 years. Somewhat
embarrassingly, the Lucerne record from the 1992 M 7.3 Landers
earthquake had motions too large to be accommodated by the
Law and Justice Center isolators. The isolator design parameters
appear to have been derived from the USGS National
Probabilistic Seismic Hazard Maps. In turn, The USGS PSHA
maps rely heavily on the NGA 2 GMPE’s to predict shaking as a
function of magnitude, distance, and site parameters.
We attempt to clarify the long-period near-source ground motions
that are predicted by the NGA 2 GMPE’s by investigating the
predictions for 10-s response spectral displacements for very
near-source locations in large crustal earthquakes. We assume
that 10-s response spectral values are approximately equal to
the peak ground displacement experienced at a site. In this case,
the 10-s spectral displacements at 0-km distance to the rupture
should be comparable to 2/3 the average slip that occurred in the
earthquake (see Aagaard, Hall, and Heaton, 2004). There are
five separate GMPE’s (developed by different research groups)
and the average 10-s, 0-km, spectral displacement is 0.88 m,
assuming M=8.0. This spectral displacement should be typical of
the very near-source region of a shallow fault with average slip
of about 1.3 m, which is far smaller than is expected from a M 8
earthquake. We discuss why the NGA 2 long-period GMPE’s are
far smaller than is expected from this simple check.
Laboratory investigation of friction along rock joints and
identification of peaks in shear stiffness prior to the joint’s
shear failure, Ahmadreza Hedayat, and John Hinton (Poster
Presentation 028)
Laboratory studies of friction along synthetic and natural fault
gauge materials have been instrumental in providing a better
understanding of the fault deformation mechanisms and when
combined with geophysical monitoring techniques, have been
very successful in assessing the real (true) contact area. The
amplitude and velocity of the waves recorded continuously
during friction experiments provide new insights into these
dynamic processes occurring along faults.
This study involved performing direct shear experiments on
gypsum and Indiana limestone rock joints with simultaneous
monitoring of the ultrasonic waves propagated through the rock
joints. The specimens were loaded in a direct shear apparatus in
which the normal stress across the joint was applied first and
then the shear load was increased with a constant displacement
rate until shear failure occurred. Two arrays, each with 13
embedded ultrasonic transducers, were placed on the sides of
the rock joint during the direct shear testing and the investigation
of failure was carried out by acquiring transmitted and reflected
shear waves during the experiment.
The contact area and its evolution during friction is a key factor
in frictional sliding, and ultrasonic waves provided an estimate of
WELCOME
2016 SCEC Annual Meeting page 186
the fracture (fault) stiffness which is affected by the true contact
area. At the interfaces between the grains, the ultrasonic wave
transmits through the contact points and reflects at air gaps.
Thus, the intensity of the wave transmitted through the rock joint
is a measure of the contact area and the elastic stiffness of the
contacting surfaces. In this study, the fracture shear stiffness is
calculated based on the ratio of the transmitted to reflected
ultrasonic waves. The fracture shear stiffness measured at
different areas along the joint showed an increased with the
application of the shear load and reached a maximum (peak)
prior to reaching the maximum shear strength of the rock joint.
Peaks in the fracture shear stiffness were systematically
observed for different areas of the rock joint and were found to
be related to the local areas of failure (damage) along the rock
joint. Identification of peaks in the fracture shear stiffness prior to
the failure of the rock joint was the major finding of this study and
this poster provides a summary of the research approach and
findings.
New Investigations on the Hollywood-Raymond Fault
Zone, Los Angeles, California, Janis Hernandez, Rufus D
Catchings, Robert R Sickler, and Mark R Goldman (Poster
Presentation 081)
South of the Transverse Ranges in southern California, the
transition zone between the Hollywood and Raymond faults has
been mapped as a series of left-stepping, E-NE-trending faults
(Weber et al., 1980). Geomorphic expression of faulting in this
area has been obscured by urbanization, and limited
geotechnical or fault investigations are available for review. As
part of our evaluation of this area for Alquist-Priolo (AP)
earthquake fault-zone mapping, we performed a review of
existing studies, aerial photos, and new trench studies required
by the City of Los Angeles. A new fault study conducted along
the western end of the Raymond Fault revealed a steeply north-
dipping, continuous fault strand that extends to within a few feet
of the ground surface. This fault exhibits down-to-the-north
displacement of Holocene marsh deposits and coincides with a
marshy wetland that existed until the early 1920’s. Borehole
cores reveal a parallel set of fault strands, constrained by
radiocarbon methods to be > 11.7 ka. Because active shallow
faulting exists at this site, a guided-wave seismic study was
conducted to test whether the Raymond fault extended along a
linear valley southwestward toward the Hollywood Fault. The
data revealed discrete zones of high peak ground velocity
amplitudes that coincide with 4 separate fault strands separated
over a distance of about 350 meters. These studies support the
hypothesis that the Raymond fault zone extends southwestward
toward the Hollywood fault, and the trenching provides positive
evidence of Holocene faulting further west along the Raymond
fault than previously known. This new information can be used in
preparation of AP earthquake fault zone maps, as well as in
regional seismic hazards assessments that consider whether the
Hollywood and Raymond faults are capable of rupturing together.
Coseismic Strengthening of the Shallow Subduction
Megathrust Further Enhances Inelastic Wedge Failure
and Efficiency of Tsunami Generation, Evan T Hirakawa,
and Shuo Ma (Poster Presentation 069)
The shallow portion (upper 5 – 10 km) of the subduction
megathrust fault, governed by velocity-strengthening friction, has
generally been considered to reduce near-trench slip and
tsunamigenesis. However, Kozdon and Dunham (2013) showed
that large shallow slip can still occur in the presence of velocity-
strengthening friction if the rupture is driven strongly from the
deeper fault segment (e.g., the 2011 Tohoku earthquake). Here
we extend this dynamic rupture model to include inelastic wedge
failure (Seno, 2000; Tanioka and Seno, 2001; Ma, 2012; Ma and
Hirakawa, 2013). We show that increased basal friction due to
velocity strengthening can further drive the outer wedge into
failure, which was first proposed by Wang and Hu (2006).
Inelastic wedge failure reduces slip on the fault, but at the same
time increases the efficiency of generating seafloor uplift (the
uplift efficiency is defined as the ratio of the volume of uplift to
seismic potency, which is dimensionless), especially when the
fault is shallowly dipping. As the transition of friction from velocity
weakening to velocity strengthening is more abrupt strain can
localize in the outer wedge developing mega-splay faults. This
model can explain a range of different mechanisms for
tsunamigenesis: large near-trench slip, splay faulting, and
distributed inelastic wedge failure. The occurrence of each is
controlled by the wedge strength and fault friction properties.
Prepared by LLNL under Contract DE-AC52-07NA27344. LLNL-
ABS-700239.
Observations from preliminary 1:24,000 scale bedrock
mapping in the western Joshua Tree region: Potential
connections between historic seismicity and evolving fault
systems, Ann Hislop, Robert E Powell, Sean P Bemis, David P
Moecher, and Luke C Sabala (Poster Presentation 102)
Bedrock mapping in the epicentral areas of the 1992 Joshua Tree
earthquake (Mw 6.1) and Landers aftershocks (Mw 5.7, 5.8)
south of Pinto Mountain Fault in southern California has further
delineated the geometry, distribution, and relative chronology of
brittle structures in the northern Little San Bernardino Mts. Of
particular interest are (1) the interactions of the Dillon Shear Zone
set of NW-trending faults with three younger N-striking dextral
faults named from west to east; Long Canyon, Wide Canyon and
Eureka Peak; and (2) the kinematic roles of the Dillon Shear
Zone and the N-S faults in relation to the San Andreas Fault.
Analysis of petrofabrics and microstructures at hand sample and
thin section scale in oriented samples, and comparison to map
scale structures for elucidation of kinematics and the collection
and analysis of remotely sensed/aerial data was used to
determine orientation of outcrop to map scale fractures and
joints, and their potential relationship to the N-striking faults
("proto-faults").
Detailed bedrock mapping indicates dextral offsets of 500m on
Long Canyon Fault, ~1300m on Eureka Peak Fault SW branch
and 180 m on the SE branch all using calc-silicate and marble
beds. A new fault (Black Rock Canyon) links Wide Canyon and
Eureka Peak Faults and has ~30 m of dextral offset. Two left
restraining steps are identified; one between Burnt Mountain and
Wide Canyon Faults occupied by Warren Peak and the other at
the orientation change between northern and southern Eureka
Peak Fault occupied by Eureka Peak. Fracture study sites
represent damage zones characterized by increasing
concentration of vertical crack sets with mm to centimeter
spacing without offsets and unconsolidated breccias.
WELCOME
2016 SCEC Annual Meeting page 187
In a traverse along the Long Canyon Fault, AFT and (U-Th)/He
thermochronologic data (Sabala, 2010) indicate an abrupt
boundary between rapid uplift within ~ 12 km of the San Andreas
Fault (SAF) and slower uplift to the north. When the boundary on
this traverse is extended southeastward along the Dillon Shear
Zone structural grain it intersects the epicenter of the 1992
Joshua Tree earthquake on the southern Eureka Peak Fault and
divides the three N-striking faults into north and south segments.
A flower structure is hypothesized along the line described above
that links to the SAF at depth and could explain the rapid uplift or
pop-up section of the Little San Bernardino Mountain
Escarpment closer to the SAF.
The 14 km epicenteral depth of the Joshua Tree earthquake on
Eureka Peak Fault coincides roughly with the 17 km projected
depth of the NE dipping SAF, supporting the hypothesis that
these N-striking faults are a reflection of tectonic linkage between
the SAF and the Eastern California Shear Zone. The N-striking
faults may be upper crustal brittle fault initiation caused by the
upward migration of a through-going break in the middle crust
and an eastern migration from the current central California
segment of the SAF, partially bypassing the Big Bend slip
impediment.
Dynamic stress triggering in the 2010-2011 Canterbury
earthquake sequence, Caroline Holden, Charles A Williams,
and David A Rhoades (Poster Presentation 195)
The Canterbury earthquake sequence started in 2010 with the
M7.1 Darfield earthquake. It was followed by over 10,000
aftershocks including the tragic M6.2 February 2011
Christchurch earthquake and the latest M5.7 Valentine
earthquake (Kaiser et al., 2016). Steacy et al. (2013) suggest that
100% of the M5.5+ aftershocks occur in regions of increased
static Coulomb stress changes. In this study we investigate the
role of dynamic stress triggering on aftershock production, with a
particular emphasis on the static and dynamic effects on smaller
aftershocks.
Based on the kinematic source models of Holden (2011, 2012),
we calculate dynamic stressgrams: expressions of the evolution
in time of the stress tensor at specific locations. We then derive
time-dependent Coulomb failure stress values and respective
peak dynamic stress values, as well as other stress-related
quantities. For comparison, we also compute static stress values
using both the same source models as the dynamic calculations,
as well as more complex source models based on geodetic
inversions. We produce interpolated maps of these metrics.
Finally we statistically compare the correlation of aftershock
production with the static and dynamic approaches. We do so for
a range of triggering times and a range of magnitudes
Fault Deformation and Segmentation of the Newport-
Inglewood Rose Canyon, and San Onofre Trend Fault
Systems from New High-Resolution 3D Seismic Imagery,
James J Holmes, Neal W Driscoll, and Graham M Kent (Poster
Presentation 107)
The Inner California Borderlands (ICB) is situated off the coast of
southern California and northern Baja. The structural and
geomorphic characteristics of the area record a middle Oligocene
transition from subduction to microplate capture along the
California coast. Marine stratigraphic evidence shows large-
scale extension and rotation overprinted by modern strike-slip
deformation. Geodetic and geologic observations indicate that
approximately 6-8 mm/yr of Pacific-North American relative plate
motion is accommodated by offshore strike-slip faulting in the
ICB.
The farthest inshore fault system, the Newport-Inglewood Rose
Canyon (NIRC) Faultis a dextral strike-slip system that is
primarily offshore for approximately 120 kmfrom San Diego to
the San Joaquin Hills near Newport Beach, California. Based on
trenching and well data, the NIRC Fault Holocene slip rate is 1.5-
2.0 mm/yr to the south and 0.5-1.0 mm/yr along its northern
extent. An earthquake rupturing the entire length of the system
could produce an Mw 7.0 earthquake or larger.
West of the main segments of the NIRC Fault is the San Onofre
Trend (SOT)along the continental slope. Previous work
concluded that this is part of a strike-slip system that eventually
merges with the NIRC Fault. Others have interpreted this system
as deformation associated with the Oceanside Blind Thrust fault
purported to underlie most of the region.
In late 2013, we acquired the first high-resolution 3D Parallel
Cable (P-Cable) seismic surveys of the NIRC and SOT faults as
part of the Southern California Regional FaultMapping project
aboard the R/V New Horizon. Analysis of these data volumes
provides important new insights and constraints on the fault
segmentation and transfer of deformation.
Based on this new data, we’ve mapped several small fault
strands associated with the SOT that we observe to link up with
a westward jog in the NIRC Fault on the shelf. Our observations
are that these strands are strike-slip features associated with a
dying splay of the NIRC system rather than compressional
features associated with a regional thrust.
Tracking Seasonal Influences on Stress Changes using
Estimates of Anomalous Geodetic Strains, Bill E Holt,
Meredith L Kraner, Adrian A Borsa, and Alireza Bahadori (Poster
Presentation 165)
Quantifying transient tectonic signals continues to reveal new
insights into fault behavior and crust/mantle rheology [Segall and
Matthews, 1997; McGuire and Segall, 2003; Freed and
Bürgmann, 2004; Ji and Herring, 2012, 2013; Mavrommatis et al,
2014]. Recent work has shown that transient signals associated
with snow, ice, and groundwater loading changes can influence
seismicity rates and potentially trigger events [Heki, 2003; Lutrell
et al., 2007; Bettinelli et al., 2008; Gonzalez et al., 2012; Amos et
al., 2014]. Here we use seasonal signals in addition to other
transients to fully quantify the time-dependent transient strain
field in southern California [Holt and Shcherbenko, 2013], along
with its statistical significance, using the PBO cGPS data.
We use the measured strain changes to resolve stress changes
on existing fault structures through time. Our analysis of the
strain response to seasonal signals is motivated by the
recognition of the importance of seasonal hydrologic loading
[Borsa et al., 2014] and the apparent stress changes that it
makes on structures [Kraner et al., 2016]. The length scales of
hydrologic loading influence stresses at significant seismogenic
depths [Argus et al., 2014, Amos et al, 2014; Borsa et al., 2014].
WELCOME
2016 SCEC Annual Meeting page 188
Thus stress changes associated with seasonal loading are not
merely surficial, but should affect the entire seismogenic layer if
seasonal signals have wavelengths of order tens of kilometers.
Thermoelastic strains [Prawirodirdjo et al., 2006; Ben Zion and
Allam, 2013] may also have an important impact on horizontal
seasonal strain and stress changes. In the case of thermoelastic
forcing, calculations show that if the near surface is coupled with
the underlying elastic layer, the depth of penetration of strain is
proportional to the length scales of the anomaly (tens of
kilometers) [Prawirodirdjo et al., 2006].
Additionally, we further develop a transient analysis method [Holt
and Shcherbenko, 2013] to incorporate seasonal signals and
investigate the influence on coulomb stress changes on faults
and model seismicity rate changes. We quantify uncertainties,
characterize the strain/stress effects on existing structures
caused by seasonal anomalies, and investigate their influence on
seismicity rate models. Here we mainly present results obtained
to date for transient seasonal strains in the regions surrounding
the South Napa earthquake.
Shear Wave Velocities in the Central Eastern United
States, Mehrdad Hosseini, Paul G Somerville, Andreas
Skarlatoudis, Jeff R Bayless, and Hong Kie Thio (Poster
Presentation 170)
USARRAY and ANSS seismic stations provide an invaluable
waveform dataset for studying ground motion attenuation in the
Central and Eastern United States. However, the dataset is
useful only after the site effects at each station are well
understood. Since there are many stations in USARRAY and
ANSS, it would be costly to accurately determine sub-surface
velocity structure beneath every station using geophysical
exploration techniques involving arrays, such as ReMi and
SASW. It would be more economical to estimate the site effects
with waveforms recorded at the seismic stations. Such
approaches have been widely applied, but most of them involve
the frequency dependent ratio between the horizontal and
vertical component of either ambient noise or S waves from
earthquakes. The horizontal component of P waves can also be
used to infer sub-surface velocity structure. It is demonstrated
that the ratio of radial to vertical P waves is mostly sensitive to
sub-surface shear velocity, so the radial/vertical ratio of the P
wave is a good indicator of subsurface shear velocity. Therefore,
the subsurface velocity structure can be estimated using an
approach similar to teleseismic P receiver functions, but at much
smaller scale and higher frequency that matches well with results
from ReMi and Refraction/Reflection techniques (Ni et al. 2014).
We used the approach introduced by Ni et al. (2014) and
calibrated our models with findings from Ni and Somerville (2013)
to obtain data on shallow shear-wave structure in the Central and
Eastern United States. A rich database of local earthquakes from
2009 to 2013 recorded at ~560 seismic stations was analyzed
and associated receiver functions were inverted. The results are
in reasonable agreement with local geology and with the results
of Ni and Somerville (2013), Ni et al. (2014) and Kim et al. (2014).
The data are made publically available for future refinement.
Were the 1952 Kern County and 1933 Long Beach,
California, Earthquakes Induced? Susan E Hough, Victor C
Tsai, Robert L Walker, Morgan T Page, and Seyed M Hosseini
(Poster Presentation 171)
Several recent studies have presented evidence that significant
induced earthquakes occurred in a number of regions during the
20th century related to either production or early wastewater
injection. We consider whether the Mw6.4 Long Beach and
Mw7.3 1952 Kern County earthquakes might have been induced
by production in the Huntington Beach and Wheeler Ridge oil
fields, respectively. The Long Beach earthquake occurred within
9 months of the start of directional drilling that first exploited
offshore tideland reserves at depths of ≈1200 m; the well location
was within ≈3 km of the event epicenter. The Kern County
earthquake occurred 111 days following the first exploitation of
deep Eocene production horizons within the Wheeler Ridge field
at depths reaching 3 km, within ≈1 km of the White Wolf fault
(WWF); the epicenter of this earthquake is poorly constrained but
the preferred epicenter is within ≈7 km of the well. While
production in the Wheeler Ridge field would have likely reduced
pore pressure, inhibiting failure on the WWF assuming a
Coulomb failure criteria, we present a model based on analytical
solutions with model parameters constrained from detailed
industry data, whereby direct pore pressure effects were blocked
by a normal fault that created an impermeable barrier close to
the WWF, allowing the normal stress change associated with
production to dominate, thereby promoting failure by unclamping
the fault. Our proposed triggering mechanism is consistent with
the observation that significant earthquakes are only rarely
induced by production in proximity to major faults. Our results
also suggest that significant induced earthquakes in southern
California during the early 20th century might have been
associated with industry practices that are no longer employed
(i.e., production without water re-injection). The occurrence of
significant earthquakes during the earthquake 20th century
therefore does not necessarily imply a high likely of induced
earthquakes at the present time.
Spatial and temporal relationships between tremor and
slip in large slow slip earthquakes on the Cascadia
subduction zone, Heidi Houston, Kelley A Hall, and David A
Schmidt (Poster Presentation 259)
Slow slip and tremor migrate in tandem along several subduction
zones at about 8 km/day in large Episodic Tremor and Slip (ETS)
events. These events occur downdip of the locked megathrust,
and thus the updip limits of slow slip or tremor provide potential
constraints on the slip budget of subduction zones and on the
downdip edge of the locked zone, which inform hazard
assessments for major cities including Seattle and Tacoma.
Inversions of GPS displacements suggest that slow slip extended
about 15 km updip of the seismically-detected tremor in the 2010
M6.8 and 2012 M6.6 ETS events (Houston, AGU abstract, 2012;
Hall and Houston, AGU abstract, 2014). In inversions where slip
is restricted to regions with tremor, the inversion forces large and
unphysical slip at the updip edge of tremor, well over that which
could accumulate during an interETS period. This also suggests
that slip extends updip of tremor, implying spatial variations of
the ratio of tremor activity to slip on the fault. While tremor
appears to have a clear updip limit, the updip extent of slow slip
is less clear. We explore the time evolution of slip during the 5-
week-long 2010 event, using the Extended Network Inversion
Filter and tremor locations from the PNSN. Our preliminary
results yield a M6.8 event that begins near Seattle and
propagates mainly to the north with some smaller slip to the
WELCOME
2016 SCEC Annual Meeting page 189
south, following the propagation of the tremor. The size and the
larger slip to the north are consistent with our static inversion
results, as well as with the tremor density. The slip pulse nature
of the ETS process is clearly imaged, with regions continuing to
slip for several days after tremor has passed through, but not for
the entire duration of the event. High concentrations of tremor
propagate along with the highest slip rates, although some
tremor can occur before the leading edge. The initial tremor
begins at a depth of ~45 km and migrates up-dip before
propagating along strike. The GPS data are not able resolve slip
until the tremor has migrated up-dip. We also find that GPS
stations closer to the initiation region of the ETS show higher
deformation rates than the stations near the termination. We
explore how variations in tremor amplitude correspond to
variations in slip (amplitudes from C. Ulberg and K. Creager). Our
goal is to determine how closely tremor activity mimics slip
activity, and understand where and why they may manifest
differently.
The aseismic slip acceleration before the 2015 Illapel Mw
8.4 event from repeating earthquake observations, Hui
Huang, and Lingsen Meng (Poster Presentation 160)
Recent studies provide evidence of accelerating short-term or
long-term seismicity possibly accompanied by aseismic slip
preceding large interplate earthquakes, such as the 2011
Tohoku-Oki Mw 9.0 earthquake and the 2014 Iquique Mw 8.2
earthquake. However, the generality of such precursory patterns
is still under heavy debate. Two underlying mechanisms are
proposed to explain the foreshock sequences: cascade
triggering of a mainshock-aftershock sequence following Omori’s
Law and the large-scale aseismic slip or other transient
processes that load the foreshocks. Although acceleration of
seismicity is observed before some large earthquakes, the
evidence of slow slip is often unclear due to the weak signals or
low resolution of geodetic observations. Repeating earthquake
analysis is an alternative way to indirectly measure aseismic slip
on the plate interface. The 2015 Illapel Mw 8.4 event ruptured at
the central part of the Chilean subduction zone. A recent study
shows that the seismicity rate within 50 km from the Illapel
mainshock elevated ~60 days before the mainshock. However, it
is not clear whether slow slip developed on the plate interface
prior to its occurrence. In this study, we investigate the repeating
earthquakes from 2011 to 2016 along the Chilean margin (18-
38°S) to infer the aseismic slip evolution along the different
segments: the 2014 Iquique rupture zone in northern Chile, the
2015 Illapel rupture zone in central Chile and the 2010 Maule
rupture zone in southern Chile. We find that the aseismic slip
around the Illapel rupture zone starts to accelerate ~150 days
before the mainshock. In contrast, around the Maule and Iquique
rupture zones, the aseismic slip rate tends to be steadily
decreasing in the years before the Illapel event. This suggests
that the area around the Illapel rupture zone experiences
accelerating slow slip during its late interseismic period. In
contrast, the decaying slow slip rate around the Maule and
Iquique rupture zones is consistent with the afterslip model. The
deformation transient around the Illapel rupture zone may
promote the unstable slip during the nucleation process of the
Illapel mainshock. The unique aseismic slip acceleration found
around the Illapel rupture zone among the whole Chilean margin
highlights the importance of monitoring slow slip for possible
hazard assessment in the locked portion of the megathrust.
Coseismic and postseismic deformation from the 2010
Mw 7.2 El Mayor-Cucapah Earthquake in Baja California:
Lithospheric structure and deformation in the Salton
Trough, Mong-Han Huang, Haylee Dickinson, Andrew Freed,
Eric J Fielding, Roland Bürgmann, and Alejandro A Gonzalez-
Ortega (Poster Presentation 157)
Coseismic and postseismic deformation from the 2010 Mw 7.2 El
Mayor-Cucapah Earthquake in Baja California: Lithospheric
structure and deformation in the Salton Trough
The 2010 Mw 7.2 El Mayor-Cucapah (EMC) Earthquake ruptured
about 120 km along several NW-striking faults to the west of the
main plate boundary fault in the Salton Trough of Baja California,
Mexico. The Cerro Prieto Fault to the east is the most active fault
at this latitude in the Pacific-North America plate boundary,
absorbing about 45 mm/year of relative motion and connecting
the San Andreas Fault and other related faults in California to the
oblique seafloor spreading in the Gulf of California. We analysed
interferometric synthetic aperture radar (InSAR), SAR and optical
pixel offsets, and continuous and campaign GPS data to develop
an EMC coseismic rupture model with 9 fault segments to fit the
complex structure of the faults that are mostly sub-parallel to the
Cerro Prieto Fault. Coseismic slip inversion with a layered elastic
model shows that largely right-lateral slip is confined to the upper
10 km with strong variations along strike. Near-field GPS enables
geodetic measurement of the slip on a north-striking normal fault
that ruptured at the beginning of the earthquake and was
previously inferred from seismic waveforms.
Postseismic deformation after the EMC Earthquake shows the
Earth’s response to the large coseismic stress changes. We use
continuous GPS measurements from the Plate Boundary
Observatory operated by UNAVCO to measure the first five years
of postseismic deformation. We find that afterslip on faults
beneath the coseismic rupture alone cannot explain the far-field
displacements. Those displacements are best explained by
viscoelastic relaxation of the lower crust and upper mantle. The
Salton Trough in southernmost California and northernmost
Mexico is an incipient rift with oblique extension of the lithosphere
and extremely high heat flow. Passive and active seismic studies
of the lithospheric structure show extremely thin crust and mantle
lid. We built a viscoelastic 3D finite element model of the
lithosphere and asthenosphere for the Salton Trough and
adjacent tectonic provinces with the EMC coseismic faults
embedded inside. An approximation of the coseismic slip was
imposed on the model, allowed to relax for 5 years, and then
compared to the observed surface deformation. Systematic
exploration of the viscoelastic parameters shows that horizontal
and vertical heterogeneity is required to fit the spatial pattern of
the postseismic deformation.
Our preferred viscoelastic structure for the region has weaker
viscosity layers beneath the Salton Trough than adjacent blocks
that are consistent with the inferred differences in the geotherms.
Defining a mechanical lithosphere as rocks that have viscosities
> 1019 Pa s (able to sustain stresses for more than 100 years),
we infer the thickness of the lithosphere beneath the Salton
Trough to be 32 km and 65 km beneath the Peninsula Ranges to
the west. These mechanical lithosphere-asthenosphere
WELCOME
2016 SCEC Annual Meeting page 190
boundaries (LABs) are shallower than the observed seismic
LABs, but probably better represent the strength of the tectonic
plates in this area.
A Flexible and Memory-Efficient Displacement-Based
Approach to Modeling Attenuation in Wave Propagation,
Md Monsurul Huda, Ricardo Taborda, and Naeem Khoshnevis
(Poster Presentation 270)
Attenuation in the form of energy losses due to internal friction is
a major factor in wave propagation and ground motion simulation
problems. These effects change the amplitude and dispersion
characteristics of seismic waves, especially over large distances
or in highly dissipative media. Attenuation is usually
characterized in terms of the quality factor, Q, and represented
by means of viscoelastic models made of springs and dashpots.
A recently introduced model, called the BKT model (after authors
Bielak, Karaoglu and Taborda), proposed the use of only three
mechanical elements in parallel, two Maxwell elements (each
made of a spring and a dashpot connected in series) and a Voigt
element (consisting of a spring and a dashpot connected in
parallel). The BKT model showed very good adherence to
intended values of constant Q = Q0 in the frequency domain, and
its implementation in the time domain required a reduced number
of memory variables compared to other methods used in
simulation. Though accurate and efficient, the implementation of
the original BKT model required a fixed pre-computed set of
auxiliary variables that had to be interpolated during running time
for arbitrary Q0 target values, and was limited to frequency-
independent Q. In this work, we expand the BKT model
introducing a more flexible implementation in which the model
parameters are set in run time using empirical expressions
formulated based on a numerical optimization of the model’s fit
with the target Q. We also extend the BKT model by adding a
third Maxwell element to improve accuracy permit the simulation
of frequency dependent Q = Q(f). The addition of a third Maxwell
increases memory usage but the model still retains its memory-
efficiency in comparative terms with respect to other alternative
implementations. More important is the fact that this expansion
allows us to model frequency dependent Q, a critical factor in
deterministic physics-based earthquake simulations done at
frequencies of engineering interest (f > 1 Hz). In this work we
present initial tests of the new BKT models with respect to semi-
analytical solutions of idealized problems 1D and 3D, including
comparisons with an equivalent frequency-wavenumber
implementation. The results show that the model behaves well.
Site Response During And After Nonlinear Soil Behavior
Has Occurred, Kenneth S Hudson (Poster Presentation 282)
Instrumented geotechnical field sites are designed to capture the
infrequent but critically important in situ case histories of ground
response, deformation, and liquefaction during significant
earthquakes that generate high intensity ground shaking and
large strains. The University of California at Santa Barbara has
been monitoring densely instrumented geotechnical array field
sites for almost three decades, with continuous recording now for
more than a decade. When seismic waves travel into soil with
sufficiently large ground motions, the soil behaves nonlinearly
meaning the shear modulus of the material decreases from the
linear value observed during weak ground motions. The
degraded shear modulus can continue to affect a site for a period
of time by changing the soil response during smaller ground
motions after the large event. Decreased shear modulus is
inferred when a decrease of shear wave velocity between two
sensors in a vertical downhole array is observed. This velocity is
calculated by measuring the difference in shear wave arrival
times between the sensors using normalized cross correlation.
The trend of decreasing shear wave velocity with increasing peak
ground acceleration is observed at multiple geotechnical array
field sites. The length of time the decreased velocity remains
following stronger shaking is analyzed using more than 450
events over more than a decade at the Wildlife Liquefaction Array
(WLA). Using both monthly and yearly velocity averages between
sensors, there is evidence that suggests the shear wave velocity
remains low over a period of months following larger significant
shaking events at the site. In addition, at WLA there is evidence
that the decrease in shear wave velocity can be detected at
ground motion levels as low as 20 cm/s2.
Tectonic tremor in the San Jacinto Fault, near the Anza
Gap, detected by multiple mini seismic arrays, Alexandra
A Hutchison, and Abhijit Ghosh (Poster Presentation 242)
Using several detection and location methods, employing data
from both network and mini high density seismic arrays, we
detect several instances of non-volcanic tremor that are possibly
located on the San Jacinto Fault (SJF) near the Anza Seismic
Gap. We apply multi-beam backprojection [Ghosh et al., 2012],
envelope cross correlation [Wech & Creager, 2008], spectral
analyses, and visual inspection to the seismic array data from
June 2011. Each location technique provides similar source
locations for each of the tremor events in our catalog. Each event
has been scrupulously tested with each of the methods, resulting
in a conservative catalog. All of the tremor windows also have a
low slowness, suggesting that they have a deeper source, ruling
out potential sources of surface noise that may result in false
detections. To further avoid such noise, we determine that the
best frequency band to detect tremor in this region is 6-8 Hz,
which ensures exclusion of the frequency band emitted by trains
(3-5 Hz) [Cristea-Planton et al., 2013] that can often pollute the
standard tremor frequency band of 2-8 Hz. These results indicate
that stress dynamics and seismicity cycles in this section of the
SJF may be affected by tremor and likely slow-slip, given their
frequent coincidence in other areas.
Validation of 3D velocity models using earthquakes with
shallow slip: case study of the Mw6.0 2014 South Napa,
California, event, Walter Imperatori, and Frantisek Gallovic
(Poster Presentation 288)
Three-dimensional velocity models constitute a key element in
strong ground motion modelling, e.g., earthquake hazard
assessment. Their validation is typically based on modelling
weak earthquakes with foci limited to depths greater than ~5km.
However, ruptures during moderate and large earthquakes can
propagate to shallower depths (and eventually reach the
surface). For such shallow sources velocity models may not be
validated with sufficient accuracy. In this respect, we conduct a
series of tests based on the Mw6.0 2014 South Napa
earthquake, which was characterised by a very shallow slip
asperity, to assess the performance of the USGS 3D San
Francisco Bay area velocity model within 20km fault distance.
Our study indicates that the velocity model performs generally
WELCOME
2016 SCEC Annual Meeting page 191
well with some exceptions, where large amplitude surface waves
not present in the observed data are systematically excited. We
exclude that more complex fault geometries or slightly deeper
slip could result in a better fit of the observed data. Contrarily, we
demonstrate that smoothing the velocity model (i.e. reducing the
strong velocity contrasts between basin fill and bedrock)
effectively attenuates the spurious oscillations.
Toward Rapid, Robust Characterization of Subduction
Zone Earthquakes: Application to the 2015 Illapel, Chile
Earthquake, Tisha Irwin, and Gareth J Funning (Poster
Presentation 159)
Accurate and robust knowledge of earthquake source
parameters is crucial for hazard assessment and disaster
response. The short revisit interval and open and rapid data
availability of the European Space Agency’s Sentinel-1 synthetic
aperture radar (SAR) mission mean that SAR interferometry
(InSAR) can be a viable method for rapid characterization of
major earthquakes, provided the tools for rapid data analysis are
available. Thrust-faulting events in subduction zones are
particularly damaging; they may generate devastating tsunami
and trigger damaging aftershocks. In this study, we develop
methods to rapidly model source parameters using InSAR data
from the September 16, 2015, Mw 8.3, Illapel, Chile earthquake.
Subduction zone fault geometries have been characterized from
seismicity and other data (e.g. the Slab 1.0 model). Since a priori
knowledge of fault geometry would eliminate one of the more
time-consuming aspects of source parameter modeling, we
begin with the Slab 1.0 model of South America. We use a
triangular fault mesh to accommodate the irregular fault
geometry, then calculate Green’s functions using analytical
solutions for triangular dislocations. We then compare an
interferogram downsampling schema using conventional
quadtree decomposition to one based on the Green’s functions
and model resolution.
Preliminary results suggest that taking advantage of these a
priori fault geometry and interferogram sampling models
produces a slip model with maximum slip ~12 m and moment
magnitude of 8.39, and which reproduces the observed surface
displacements with a root mean square misfit of ~6 mm. Pre-
event calculation of Green’s functions and interferogram
sampling schemes, combined with predetermined fault geometry
and short satellite revisit intervals, will allow InSAR to play a
complementary role to seismic waveform analysis in the rapid
characterization of earthquakes.
The bridge from earthquake geology to earthquake
seismology, David D Jackson (Oral Presentation 9/12/16
11:00)
Recurrence intervals of large earthquakes exceed the
instrumental earthquake record, so we rely on paleo-seismic and
surface deformation evidence to infer long-term earthquake rates
and recurrence statistics. Paleo-seismic studies generally imply
quasi-periodic recurrence and greater rates for large
earthquakes than those deduced from instrumental data coupled
with Gutenberg-Richter magnitude relations.
Under SCEC leadership a team of paleo-seismic experts
compiled the most reliable data available for use in the third
Uniform California Earthquake Rupture Forecast (UCERF3).
They reported dates of observed displacements at 32 sites on 13
named faults in California. Corrected for multiple-site ruptures,
the total reported paleo-event rate at those sites exceeds about
4 per century, yet none has occurred since 1916. Such a long
hiatus is extremely unlikely for a Poisson process and even less
probable for an ensemble of quasi-period processes.
Hypotheses for the hiatus of paleo-events include (1) extreme
luck, (2) unexplained regional fault interaction, or (3) mistaken
identification of non-seismic displacements as evidence of large
earthquakes, or counting multiple branches of single ruptures as
separate ruptures. The first can be rejected with 99% confidence.
There is no evidence for the second in the pre-1916 paleo-
seismic history or in any theoretical models yet reported.
Temporal clustering of large quakes (“supercycles”?) could in
principle explain the hiatus, but individual site records and rate-
state frictional models suggest quasi-periodic recurrence
instead. The third hypothesis could explain the observed
quiescence because mistaken identity would be prevented by
instrumental seismic data during the last century. In any case the
paleo-event hiatus poses a serious challenge to earthquake
forecast models. It also begs the larger question of how 2-
dimensional fault-based observations and models can be
reconciled with 3-D instrumental earthquake observations.
New GPS Site Velocities in San Gorgonio Pass, Southern
California and Preliminary Elastic Modeling, Naomi Jahan,
Matthew Peterson, Sally F McGill, and Joshua C Spinler (Poster
Presentation 149)
Campaign GPS observations have been collected from 23 new
sites in San Gorgonio Pass 1-3 times per year since 2013. After
3 years of observations, consistent velocities are now available
for these sites and are well constrained enough to support elastic
modeling of fault slip rates. We conducted one-dimensional
elastic modeling of right-lateral fault slip rates in two transects
across the plate boundary passing through the San Gorgonio
Pass and Desert Hot Springs. For the San Gorgonio Pass
transect, the best-fitting model puts all of the slip on the Johnson
Valley and San Jacinto faults, with a very small amount of slip on
the Pisgah-Bullion fault and no slip on either strand of the San
Andreas fault. Other models, which have almost as low a chi2,
put 9 mm/yr on the two strands of the San Andreas fault
combined. The maximum amount of slip that can be given to the
two strands of the San Andreas combined while still fitting the
observed GPS velocity profile reasonably well is 19 mm/yr. The
models indicate that the San Andreas fault can reasonably have
between 0 and 19 mm/yr of slip and the San Jacinto fault can
have between 8 and 21.5 mm/yr of slip. For the Desert Hot
Springs transect, the best-fitting model placed large amounts of
slip on the Burnt Mountain (11.4 mm/yr) and San Jacinto faults
(13.4 mm/yr), with 5.6 mm/yr of slip on the San Andreas fault
(Banning strand). Other models resulted in a nearly equally good
fit, with a San Andreas fault slip rate of 10 mm/yr, and slip on the
Burnt Mountain fault of 9 mm/yr. The maximum slip rate of the
San Andreas fault that still fits the data reasonably well is 25
mm/yr. Our results suggest that little or no elastic bending is
presently occurring across the Banning or Mission Creek faults
within the San Gorgonio Pass or Desert Hot Springs transects.
These results support interpretations in which most of the slip on
the Mojave section of the San Andreas fault transfers southward
WELCOME
2016 SCEC Annual Meeting page 192
onto the San Jacinto fault and most of the slip on the
southernmost San Andreas fault transfers northward into the
Eastern California shear zone, leaving only a small amount of slip
on the San Andreas fault between Cajon Pass and Thousand
Palms.
The East Shoreline strand of the San Andreas Fault and
its implications for the next Big One in southern California,
Susanne U Janecke, Daniel Markowski, Roger Bilham, James P
Evans, Michael Bunds, Jack Wells, Jeremy Andreini, and Robert
Quinn (Poster Presentation 116)
The southern San Andreas Fault last ruptured about 1670 AD,
and its southern tip could nucleate the next Big One (Shakeout
summary). We use new geologic data along a 15-km long swath
at the southernmost tip of the San Andreas Fault zone (SAFZ) to
determine the presence of a new important fault strand and to
evaluate its earthquake potential. Geologic mapping, structural
and stratigraphic studies, interpretation of LiDAR, false and
natural color aerial photography, published InSAR
interferograms from several time intervals, measurements from
creepmeters, published reflection and refraction seismic, and
gravity data, and preliminary data from high-resolution aerial
imagery and derivative digital elevation models acquired by
unmanned aerial vehicle, all document the presence of a second
significant strand of the San Andreas Fault zone. We named this
strand of the SAFZ the East Shoreline strand of the SAF
(ESSSAF) for its position along the northeast shore of the Salton
Sea, 1 to 3 km southwest of the main trace of the SAF.
Seven key data sets provide evidence for the existence of the
East Shoreline strand of the San Andreas Fault zone. These are:
Mapping of many tens of steeply to moderately dipping,
northwest-striking dextral faults within the East Shoreline fault
zone; Miocene (?) to Pleistocene stratigraphic units that are cut
out by these dextral faults; Monoclines, synclines and anticlines
that parallel dextral strike-slip faults in the ESSSAF for significant
distances; The dozens of left- and right -lateral cross faults that
connect between the ESSSAF and the main strand of the SAF
and form a coordinated ladder-like fault mesh; Many of the cross
faults change their strike toward parallelism near the two master
dextral faults and display a symmetric sigmoidal geometry; There
are east-striking left-lateral strike-slip faults between the
ESSSAF and the main strand of the SAF, that rotated about 45°
clockwise in a bookshelf manner between two dextral strands of
the SAF since their inception in the Pleistocene, and a close
correspondence between the mapped East Shoreline zone and
deformation on InSAR scenes.
The ESSSAF is an active dextral-reverse structure that probably
creeps at shallow crustal levels. Analysis of several sets of
InSAR interferograms show that ~7 ± 2 mm/yr of right-lateral
creep and a few mm/yr of southwest-down vertical motion
accumulated across the ESSSAF and the inactive faults of the
northern Salton Sea between 2003 and 2010. Most of this
deformation likely accumulated across the ESSSAF because
published reflection seismic studies identified mostly
homoclinally dipping beds in the northern Salton Sea.
Pleistocene, Holocene, and modern sediment are folded and
faulted within the fault zones, and older units are more deformed
than younger ones. There are linear creep-related fractures,
drought-enhanced fractures and aligned and elongate desert
sinks (FDEFS) along strike of one another in modern beach
deposits and on the floors of washes continuously along a 7 km
long stretch of the ESSSAF. Spot checking and aerial imagery
reveal more discontinuous linear fractures along ~ 18 km of the
fault zone. The main strand of the San Andreas Fault and its
immediate damage zone also has FDEFS, they closely resemble
those within the ESSSAF, and formed in well known as well as
many newly identified faults in the southern half of Durmid Hill.
Fault-related structures within the ESSSAF formed progressively
and show less deformation in the upper Brawley Formation than
in older Brawley Formation across angular unconformities.
Together the main and East Shoreline strands of the San
Andreas Fault form a coordinated sheared ladder-like fault zone
in map view, and bound dozens of left- and right-lateral cross
faults and folds between them. This active fault mesh resembles
the Brawley seismic zone immediately to the south, except that
the Durmid ladder structure of the SAFZ is a transpressional
dextral-reverse fault zone with internal uplift, whereas the
Brawley seismic zone is transtensional belt with internal
subsidence. From the Salton Sea northward, a part of the
ESSSAF may cross the Coachella Valley to Palm Springs and
San Gorgonio Pass, as we proposed in 2013. A NW branch may
become the blind (?) Palm Springs dextral-reverse fault beneath
the greater Palm Springs metropolitan area.
Durmid Hill was severely shaken by the El Major Earthquake in
April 2010, and this shaking changed the prevailing slip regime,
creep rates and locally even the slip sense at the Ferrum, Salt
Creek and Durmid creepmeters. For three years another strand
of the SAFZ became active, producing highly unusual creep
meter records along its main strand. Analysis of the data is
ongoing, but initial analysis suggests that creep along the East
Shoreline strand may partly explain these changes.
Contraction across strike-slip fault zones tends to inhibit the
nucleation of large earthquakes in favor of more neutral or
extensional stretches. There is much contraction in the Durmid
Hill ladder structure, and neutral to mildly extensional zones
farther north along the SAF may be more likely to nucleate future
large earthquakes than the southern tip of the fault zone.
The northeast branch of the ESSSAF spans at least 48 km
between Bombay Beach and the latitude of Thermal Canyon, CA.
This is constrained by published seismic data of Fuis et al.
(2012), bathymetry, published InSAR data, fault scarps, and a 14
km long belt of uplifted Pleistocene basin fill on the southwest
side of the main strand of the San Andreas fault in the Mecca
Hills. From Salt Creek to the latitude of Coachella the ESSSAF
accommodated as much as 3-7 mm/yr of right lateral creep
between 2006 and 2010 according to the InSAR data of Lindsey
et al (2014), as long as the ESSSAF is the only active fault zone
on the floor of Coachella Valley.
Site amplification effects at Heathcote Valley, New
Zealand, during the 2010-2011 Canterbury earthquakes,
Seokho Jeong, and Brendon A Bradley (Poster Presentation
283)
This poster presents a quantitative case study on the role of near
surface site effects on the ground motion intensities at Heathcote
Valley school station (HVSC), New Zealand, during the 2010-
2011 Canterbury earthquake sequence, accounting for the
WELCOME
2016 SCEC Annual Meeting page 193
surface topography, the realistic geological structure, and the
non-linear constitutive behaviour of soils.
Simulations show that ground motions at HVSC are amplified in
a wide band of frequencies, and the simulated ground motion
amplitude in high frequencies (i.e. higher than the site
fundamental frequency) is sensitive to the various modelling
assumptions.
The result demonstrates that 1D simulations without proper
consideration of the basin structure tend to underestimate the
site response, likely due to its inability to model surface waves
caused by the inclined soil-rock interface. Comparison of linear-
elastic and non-linear analyses suggests that the non-linear
hysteretic response of soils dissipates significant energy for
strong shaking events. Adoption of a simpler depth-independent
soil model revealed that such a minor simplification can lead to a
substantial underestimation of ground motion intensity.
Simulations also suggest that the pressure-dependent (no
tension) nonlinear constitutive behaviour of soils, combined with
energetic near-field vertical motions, might be the primary
mechanism of strong asymmetrical acceleration pulses observed
during the February 2011 Christchurch earthquake.
Slip history of the 2016 Mw 7.0 Kumamoto earthquake:
intra-plate rupture in complex tectonic environment, Chen
Ji, Jinglai Hao, and Zhenxing Yao (Poster Presentation 235)
We have selected 14 3-component strong motion observations,
combining with waveforms of teleseismic broadband body waves
and long period surface waves of the 2016 Kumamoto
mainshock to constrain its temporal and spatial distribution of
slip. Our preferred model is composed of Hinagu (dipping 73o)
and Futagawa (dipping 60o) fault segments. Rupture initiated at
JMA hypocenter on the Hinagu segment in dominant right-lateral
strike-slip motion. The significant rupture on the Futagawa
segment occurred 4-5 s later at a depth about 10 km with pure
strike-slip motion but normal fault component increases as the
rupture propagates close to the Aso Volcano. The total rupture
duration is 15 s. The best double couple solution of the
cumulative moment tensor of this rupture has strike, dip, rake
angles of 224o, 64o and -152o, respectively, agreeing
remarkably with the long period best double couple solutions of
USGS W-Phase solution (224o, 66o, -152o) and the GCMT
project (222o, 77o, -163o). However, the CLVD component is not
significant, consistent with USGS solution. It is noteworthy that
unlike Hinagu segment, the focal mechanisms of Mw>5
foreshocks all have sub-vertical nodal planes and slightly dipping
to the southeast, suggesting the Hinagu surface trace associate
with two fault planes in depth. This scenario is closely analogous
with the relation of 1989 Loma Prieta, earthquake and San
Andreas fault. The relationships among the slip distribution,
thermal structure near Volcano and background tectonic loading
will be addressed.
Reconciling seismicity and geodetic locking depths on the
Anza segment of the San Jacinto Fault, Junle Jiang, and
Yuri Fialko (Poster Presentation 066)
The depth extent of fault locking can be independently estimated
from seismic and geodetic observations. In Southern California,
available observations for the San Andreas and San Jacinto
Faults suggest a general agreement between the two estimates.
Among a few notable exceptions is the Anza segment of the San
Jacinto Fault, where geodetically-estimated locking depth is ~10
km while seismicity occurs as deep as 15-18 km. The
discrepancy between the seismic and geodetic estimates is at
odds with models of faults governed by the rate-state friction with
laboratory-based, depth-dependent properties, because such
models predict that the effective geodetic locking depth should
be close to or deeper than the depth of concentrated seismicity
(Jiang and Lapusta, 2015).
We suggest that the discrepancy may be due to a broad
transition zone of highly heterogeneous frictional properties,
which is not envisioned in models of depth-dependent fault
properties based on laboratory data. Using numerical models of
faults with stochastic heterogeneity in rate-state frictional
properties, we explore the interplay of diverse fault slip
phenomena in the long-term fault slip history. Our models
successfully reproduce the observed depth relation between
seismicity and geodetic locking. Furthermore, we find that the
heterogeneous transition zone introduces a great variability in
microseismicity, thus also explaining the scarcity of repeating
earthquakes in the region of Anza. Such rheological transition
may reflect the combined effects of complex structures and
properties of the deep fault zone, and probably applies to other
similar fault segments. Importantly, our models demonstrate that
these faults are capable of hosting large earthquakes rupturing
to the greatest extent of seismicity, with increased complexity in
rupture fronts downdip. Our simulations predict that spontaneous
and triggered aseismic transients should be ubiquitous in the
transition zone between large events on the Anza segment,
which can be potentially verified using high-precision geodetic
observations.
Seasonal water storage modulating seismicity on
California faults, Christopher W Johnson, Yuning Fu, and
Roland Bürgmann (Poster Presentation 012)
In California the accumulation of winter snowpack in the Sierra
Nevada, surface water in lakes and reservoirs, and groundwater
in sedimentary basins follow the annual cycle of wet winters and
dry summers. The surface loads resulting from the seasonal
changes in water storage produce elastic deformation of the
Earth’s crust. Micro-earthquakes in California appear to follow a
subtle annual cycle, possibly in response to the water load.
Previous studies posit that temperature, atmospheric pressure,
or hydrologic changes may strain the lithosphere and promote
additional earthquakes above background levels. Here we use
GPS vertical time series (2006 – 2015) to constrain models of
monthly hydrospheric loading and compute annual peak-to-peak
stresses on faults throughout northern California, which can
exceed 1kPa. Depending on fault geometry the addition or
removal of water increases the Coulomb failure stress. The
largest stress amplitudes are occurring on dipping reverse faults
in the Coast Ranges and along the eastern Sierra Nevada range
front. We analyze M≥2.0 earthquakes with known focal
mechanisms in northern and central California to resolve fault
normal and shear stresses for the focal geometry. Our results
reveal more earthquakes occurring during slip-encouraging
stress conditions and suggest that earthquake populations are
modulated at periods of natural loading cycles, which promote
failure by subtle stress changes. The most notable shear-stress
change occurs on more shallowly dipping structures. However,
WELCOME
2016 SCEC Annual Meeting page 194
vertically dipping strike-slip faults are common throughout
California and experience smaller amplitude stress change but
still exhibit positive correlation with seasonal loading cycles. Our
seismicity analysis suggests the annual hydrologic cycle is a
viable mechanism to promote earthquakes and provides new
insight to fault mechanical properties.
Earthquake Number Forecasts Testing, Yan Y Kagan, and
David D Jackson (Poster Presentation 301)
We study the distributions of earthquake numbers in two global
earthquake catalogs: Global Centroid-Moment Tensor (GCMT)
and Preliminary Determinations of Epicenters (Monthly Listing)
(PDE). The properties of these distributions are especially
needed to develop the number test for our forecasts of future
seismic activity rate, organized by the Collaboratory for Study of
Earthquake Predictability (CSEP). A common assumption, as
used in the CSEP tests, is that the numbers are described by the
Poisson distribution. However it is clear that the Poisson
assumption for the earthquake number distribution is incorrect,
especially for the catalogs with a lower magnitude threshold. In
contrast to the one-parameter Poisson distribution so widely
used to describe earthquake occurrence, the negative-binomial
distribution (NBD) has two parameters. The second parameter
can be used to characterize the clustering or over-dispersion of
a process. We investigate the dependence of parameters for
both distributions on the catalog magnitude threshold and the
time intervals catalog duration is subdivided. Firstly, we study
whether the Poisson law can be statistically rejected for various
catalog subdivisions. We find that for most cases of interest the
Poisson distribution can be shown to be rejected statistically at a
high significance level in favor of the NBD. Thereafter we
investigate whether these distributions fit observed distributions
of seismicity. For this purpose we study upper statistical
moments of earthquake numbers (skewness and kurtosis) and
compare them to the theoretical values for both distributions.
Empirical values for skewness/kurtosis increase for the smaller
magnitude threshold and increase with even greater intensity for
small temporal subdivision of catalogs. As is known, the Poisson
distribution for large rate values approaches the Gaussian law,
therefore its skewness/kurtosis tend both to zero for large
earthquake rates: for the Gaussian law these values are
identically zero. A calculation of the NBD skewness/kurtosis
levels based on the values of the first two statistical moments of
the distribution, shows rapid increase of these upper moments
levels. However, the observed values of skewness/kurtosis are
rising even faster. This means that for small time intervals the
earthquake number distribution is even more heavy-tailed than
the NBD expects.
Near source ground-motion modelling of the Canterbury
aftershocks and implications for engineering metrics,
Anna E Kaiser, and Caroline Holden (Poster Presentation 287)
The 2010-2011 Canterbury earthquake sequence includes the
22 February 2011 Christchurch aftershock (Mw 6.2), the June
2011 Mw 6.0 aftershock, the December 2011 magnitude (Ml) 5.8
and 6.0 aftershocks and the February 2016 Mw5.7 Valentine’s
day earthquake. These events caused widespread liquefaction,
rockfalls and heavy building damage and collapse.
Holden and Kaiser (2016) compute broadband synthetic
seismograms for the largest M5.7+ aftershocks of the 2010-2011
Canterbury earthquake sequence at near-source sites. They
employ a stochastic approach to compute the seismograms that
is not only controlled by detailed source models but also by
regional parameters derived using spectral inversion of the
extensive Canterbury strong motion dataset. Their results show
that using appropriate stress drop value and site-specific
amplification functions helps greatly to reproduce key
engineering parameters such as peak ground accelerations
(PGAs) and response spectra.
We propose to test the synthetic ground motion models of Holden
and Kaiser (2016) against newly developed metrics by Rezaeian
et al. (2015). These new metrics are ideally suited for the
validation of ground motion models for structural and
geotechnical engineering systems.
They consist of 3 time dependant metrics and 6 scalar
parameters. Results from these tests are expected to bring new
insights on the ground motion outputs in particular and valuable
information on the modelling technique in general.
Site-specific earthquake hazard characterization for
Dhaka City, Bangladesh, Mir F Karim, Zillur M Rahman,
Maksud Kamal, and Sumi Siddiqua (Poster Presentation 091)
Seismic hazard characterization is the foremost module for
earthquake risk management in a seismically vulnerable region.
The mega city Dhaka in Bangladesh is considered by many
researchers as one of the riskiest cities in the world due to many
non-engineered construction practices and poorly studied
tectonic boundary conditions. The city is built on a Plio-
Pleistocene terrace located within the subsiding Bengal basin.
The records of historical earthquakes indicate that three large
magnitude earthquakes occurred during the last 150 years within
and in close proximity to Bangladesh. Magnitudes of these
earthquakes ranged from 6.9 to 8.7 occurring between 1885 and
1918. These events caused moderate damage to buildings and
other infrastructures in Bangladesh, but the damage in Dhaka
City was negligible. It is believed that the 6.9 magnitude Bengal
earthquake occurred approximately 50 km from the city, although
there are multiple controversies about the location of the
epicenter. Many consider that the epicenter of this earthquake
was 170 km away from Dhaka City and others inferred the
epicenter to be somewhere along Madhupur fault, approximately
50 km away. The 1885 Bengal, 1897 Great Indian, and 1918
Srimangal Earthquakes are considered as the seismic sources
for site-specific seismic hazard characterization. The peak
ground acceleration (PGA), peak ground velocity (PGV), spectral
accelerations (SA) of different periods have been calculated at
the ground surface based on recently developed ground motion
prediction equations and site amplification factors. The
amplification factors are predicted from the average shear wave
velocity to a depth of 30 m (Vs30), which are estimated using
various geophysical and geotechnical investigations. The study
reveals that the city is built on very firm ground where seismic
risks are manageable provided the engineering structures
adhere to the norms of seismic regulations and building codes.
Vital Signs of the Planet: Southern California Education
Contribute to Crustal Deformation Studies Within San
Bernardino and Riverside Counties, Dan Keck (Poster
Presentation 150)
WELCOME
2016 SCEC Annual Meeting page 195
In conjunction with California State University, San Bernardino,
Inland Empire middle school, high school, and community
college teachers have used GPS to monitor movement along the
San Andreas and San Jacinto faults within the Inland Empire,
San Bernardino Mountains, and high desert regions of Southern
California since 2002. Stations observed in 2016 were selected
from those that previously had relatively poorly constrained time
series, so as to contribute useful new velocity constraints for use
by the SCEC community and others. Procedures for the study
included setting up a tripod (or spike mount), and an antenna and
receiver over existing survey monuments for an 8 hour period
each day for 3 days. GPS data were processed at the University
of Arizona using GAMIT-GLOBK and benchmark positions were
compared to those in previous years. Time series graphs were
used to estimate the north, east and vertical velocities of each
site.
Velocities for our sites were combined with velocities from
SCEC’s Crustal Motion Model version 4 (Shen et al., 2011) and
with velocities from continuous GPS stations archived at the
Plate Boundary Observatory. One-dimensional elastic modeling
of the combined data set was used to infer fault slip rates within
a transect across the plate boundary through the San Bernardino
Mountains.
Results indicate that the combined slip rate of the 15 faults within
our transect is 46 mm/yr. The San Andreas (SAF) and San
Jacinto (SJF) faults have the highest rate of movement with a
combined slip rate of 16.5-18.25 mm/yr. This is substantially less
than the published 35 mm/yr slip rate of the SAF alone in central
California. Nonetheless our inferred slip rates for the SAF and
SJF are consistent with previously published slip rates for these
faults in southern California over late Quaternary time scales.
The two faults are so close together within our transect that their
slip rates are strongly inversely related. Our best-fitting model
apportions the SJF slipping 9.75 mm/yr and the SAF slipping 8.5
mm/yr, but models with SJF slipping 15 mm/yr (model 3) and
SAF slipping 2 mm/yr, or SJF slipping 3.25 mm/yr and the SAF
slipping 14.5 mm/yr (model 2) also fit the observed site velocities
relatively well.
Compaction-induced elevated pore pressure and creep
pulsing in California faults, Mostafa Khoshmanesh, and
Manoochehr Shirzaei (Poster Presentation 162)
The creeping segment of San Andreas Fault (CSAF) is
recognized as a weak fault, namely, cannot sustain large
earthquake stress drops. Moreover, variable creep rate
constrained using kinematic models of geodetic and seismic data
implies that the fault frictional strength is both spatially and
temporally variable. Intrinsic low friction of fault zone material and
locally elevated pore pressure due to ascend of mantle-derived
fluid are proposed as possible justifications for CSAF weakness.
However, lack of plausible explanation for creep pulsing
observed at seismogenic zone in both hypotheses, calls for
rethinking of the underlying mechanisms and processes
governing the CSAF behavior. Here we provide evidence for the
role of pore pressure variation in changing the fault frictional
strength, not primarily due to mantle fluids. Using a rate- and
state-dependent friction model, we estimate fault frictional
properties between 2003 and 2011, and link their apparent
temporal variations to undulation of effective normal stress. Since
there is no evidence that tectonic stressing rate varies during this
study period, we conclude that the variation of effective normal
stress is a result of pore pressure change in the fault zone. We
show that temporally variable pore pressure and its inferred
spatial heterogeneity correlate perfectly with the variation of
surface creep rate obtained using InSAR observations.
Furthermore, our analysis of microseismicity suggests that the
temporal variation of Gutenberg-Richter b-value and released
seismic moment has respectively positive and negative
correlation with the pore pressure variations. Our results highlight
the role of 3D seal-bounded compartments formed through the
compaction of intergranular pore spaces, leading to spatially
heterogeneous elevated pore pressure and initiation of
accelerated creep events. Frictional dilation due to creep
acceleration, on the other hand, causes redistribution and
reduction of the pore pressure, completing a cycle of creep
pulsing. Applying this concept to the creep and microseismicity
data along Hayward Fault for period 1995 - 2011, yields similar
results, which suggests the observed mechanism is likely a
common explanation for creep pulsing in California faults.
Analysis of Q-factor’s parameters of Los Angeles through
Simulation and Artificial Intelligence, Naeem Khoshnevis,
and Ricardo Taborda (Poster Presentation 269)
The accurate solution of wave propagation problems requires the
appropriate representation of energy losses due to internal
friction. These losses are important because their
mischaracterization can lead to over- or under-estimation of
amplitudes and duration of ground motions. Recent studies show
that synthetics from physics-based simulations tend to attenuate
at different rates than observations, suggesting that current
modeling approaches need to be revised. In physic-based
simulation, attenuation is commonly introduced by means of
viscoelastic models. Internally, the properties of these models
are set based on the material’s quality factor, Q. The value of Q
for shear waves, Qs, is usually defined according to empirical
rules that depend on the shear wave velocity, Vs. Typical Qs-Vs
relationships are (piecewise) linear or polynomial functions.
Several relationships have been tried in the past, but there is no
consensus about what is the most appropriate one. Identifying
the parameters that define these relationships through the
solution of inverse problems is non-trivial, requires considerable
computational resources, and may not lead to a unique solution.
In this study we investigate the effectiveness of an approach that
combines the use of an artificial neural network (ANN) and a
genetic algorithm (GA) to predict ground motion amplitudes and
identify the optimal parameters in Qs-Vs relationships. We use
direct S-waves as a proxy to evaluate attenuation over distance.
First, we train the ANN to predict peak S-wave amplitudes based
on a series of simulations with an ample selection of parameters.
Then, we use the ANN as a fitness function of the GA, and use
the latter to find the optimal parameters based on comparisons
between predictions and data. As a proof-of-concept, we test the
performance of the proposed approach for the case of the 2008
Mw 5.4 Chino Hills, California, earthquake. Using this approach
we find the parameters that lead to the best fit with data. The
results include the mean and standard deviation of Qs for each
value of Vs. Our initial results for a limited number of stations
scattered throughout the simulation domain and for numerical
models with a maximum frequency of 0.5 Hz show good promise.
WELCOME
2016 SCEC Annual Meeting page 196
We recognize, however, that Q parameters may depend on
additional factors such as frequency, depth, path and site effects,
and the nature of the traveling waves. Future work will test the
method for multiple events and higher frequencies.
Music of Earthquakes, Debi Kilb, and Daniel T Trugman
(Poster Presentation 326)
A new middle school class titled 'The Music of Earthquakes' was
offered this summer as part of the Sally Ride Science Junior
Academy for Girls (https://sallyridescience.com). Exploring the
interplay between music and earthquakes, this class was part of
a new drive to include arts within the science field, a program with
the acronym STEAM, which stands for Science, Technology,
Engineering, Arts and Math. This week long, 15-hour summer
class included: playacting to understand seismic intensity
metrics, the creation of guitar pick necklaces and rubber-band
guitars, learning the difference between high- and low-
frequencies using musical instruments, understanding the
distinction between P- and S-waves using Slinkies, and
extending these concepts to understand earthquake early
warning. To practice earthquake safety, students could yell
EARTHQUAKE at an unexpected time, at which point all quickly
secured themselves using the duck-cover-hold-on technique; it
wasn't until the proclaimed passage of the three seismic phases
(P, S and Surface waves) and an 'all clear' that students ventured
back to their desk. Students were also given the opportunity to
'conduct' their own seismograms using a music stand and an
official music conductor’s baton. Students learned how to read
and understand a spectrogram using the Audacity Software
(https://sallyridescience.com), exploring variations in dynamic
range using different instrument types (in addition to screaming
at the top of their lungs), Rounding out the class, there were a
smattering of 'dance parties' to get our wiggles-on and get our
wiggles-out, fostering a genuine excitement for music and its
deep connection to wave physics and earthquake seismology.
Offshore Pacific-North America lithospheric structure and
Tohoku tsunami observations from a southern California
ocean bottom seismometer experiment, Monica D Kohler
(Oral Presentation 9/13/16 11:00)
The motivation for the offshore ALBACORE seismic experiment
was to identify the physical properties and deformation styles of
the western half of the Pacific-North America plate boundary in
southern California. An array of 34 ocean bottom seismometers
(OBSs) was deployed in 2010-2011, extending from the eastern
California Borderland into Pacific oceanic plate 300 km west of
the Patton Escarpment. The ALBACORE OBSs, together with 65
stations of the onshore Southern California Seismic Network,
were used to measure ambient noise correlation functions and
Rayleigh-wave dispersion curves which were inverted for 3D
shear-wave velocities. The resulting shear-wave velocity model
illustrates plate boundary deformation including both thickening
and thinning of the crustal and mantle lithosphere within the
California Borderland and at the westernmost edge of the North
American continent. The velocity model defines the transition
from continental lithosphere to oceanic tectonic environment,
and indicates the persistence of uppermost mantle volcanic
processes associated with East Pacific Rise spreading adjacent
to the Patton Escarpment. One of the most prominent of these
seismic structures is a low-velocity anomaly underlying San Juan
Seamount, suggesting ponding of magma at the base of the
crust, resulting in thickening and ongoing adjustment of the
lithosphere due to the localized loading. Complementary to this,
mapping of two active transpressional fault zones in the
California Borderland - the Santa Cruz-Catalina Ridge fault and
the Ferrelo fault - was carried out to characterize their geometries
using over 4500 line-km of new multibeam bathymetry data
recorded during the OBS deployment cruise, combined with
existing data. The geometry of the fault systems shows evidence
of multiple segments that could experience through-going
rupture over distances exceeding 100 km. Transpression on
west- and northwest-trending structures persists as far as 270
km south of the Transverse Ranges, extension persists in the
southern Borderland, and these faults show potential for dip-slip
rupture.
The 2011 Tohoku tsunami was serendipitously recorded by the
ALBACORE seafloor co-located pressure gauges,
demonstrating how dense array data can illustrate and validate
predictions of linear tsunami wave propagation characteristics.
Phase and group travel times were measured for the first arrival
in the pressure gauge tsunami waveforms filtered in narrow
period bands between 200 and 3000 s. For each period, phase
velocities were estimated across the pressure gauge array based
on the phase travel-time gradient using eikonal tomography.
Clear correlation is observed between the phase velocity and
long-wavelength bathymetry variations. In the deep open ocean,
phase velocity dispersion is observed for both short and long
periods. The pressure gauge tsunami records across the entire
array show multiple, large-amplitude, coherent phases arriving
one hour to more than 36 hours after the initial tsunami phase.
Beamforming and back-projection analysis of the tsunami
waveforms reveals locations of the bathymetric features in the
Pacific Ocean that contributed to the scattered, secondary
tsunami arrivals in southern California.
MyShake: Initial Observations from a Global Smartphone
Seismic Network, Qingkai Kong, Richard M Allen, and Louis
Schreier (Poster Presentation 249)
MyShake is a project aims to build a global smartphone seismic
network that harnesses the power of crowdsourcing. This poster
will present the initial observations from MyShake network after
being released to public since Feb 2016. Based on the data
collected, it is concluded that: (1) Comparing with traditional
seismic network, MyShake has the advantage to expand globally
in a short time; (2) MyShake can record earthquakes ranging
from small to large magnitude (M2.5 – M7.8 we recorded so far),
and the data recorded by MyShake will provide an independent
seismic dataset for many potential types of scientific research;
(3) MyShake can also provide seismic data where for some
places few seismic stations installed, and give additional
information about some earthquakes. Accordingly, this poster will
show the potential of using MyShake as a complementary to the
current seismic network.
Localization and instability in sheared granular materials:
Role of friction and vibration, Konik R Kothari, and Ahmed E
Elbanna (Poster Presentation 040)
Shear banding and stick-slip instabilities have been long
observed in sheared granular materials. Yet, their microscopic
underpinnings, interdependencies and variability under different
WELCOME
2016 SCEC Annual Meeting page 197
loading conditions have not been fully explored. Here, we use a
non-equilibrium thermodynamics model, the Shear
Transformation Zone theory, to investigate the dynamics of strain
localization and its connection to stability of sliding in sheared,
dry, granular materials. We consider frictional and frictionless
grains as well as presence and absence of acoustic vibrations.
Our results suggest that at low and intermediate strain rates,
persistent shear bands develop only in the absence of vibrations.
Vibrations tend to fluidize the granular network and de-localize
slip at these rates. Stick-slip is only observed for frictional grains
and it is confined to the shear band. At high strain rates, stick-slip
disappears and the different systems exhibit similar stress-slip
response. Changing the vibration intensity, duration or time of
application alters the system response and may cause long-
lasting rheological changes. We analyse these observations in
terms of possible transitions between rate strengthening and rate
weakening response facilitated by a competition between shear
induced dilation and vibration induced compaction. We discuss
the implications of our results on dynamic triggering, quiescence
and strength evolution in gouge filled fault zones.
Using GPS Imaging to Resolve Seasonal Strain in Central
California, Meredith L Kraner, William C Hammond, Corné W
Kreemer, and Geoffrey Blewitt (Poster Presentation 166)
Recently, studies using data from modern GPS have revealed
that crustal deformation can be influenced by seasonal and
nontectonic factors such as temperature, hydrology, and
atmospheric loads. Statistically significant variations in annual
micro-seismicity have also been observed. Here we build on this
new idea by developing a robust seasonal strain model of Central
California and run a series of verification routines to validate our
result. We study an 8-year period (2008-2016) such that we can
additionally test whether drought had an effect on influencing the
nontectonic seasonal signal.
We use monthly median positions of horizontal GPS
measurements to build our strain model. These positions are
detrended using robust MIDAS velocities (product of the Nevada
Geodetic Laboratory), destepped using a Heavyside function,
and demeaned to center the time series around zero. The
stations we use are carefully selected using the vertical rate of
the GPS time series during the drought period. This selection
method allows us to exclude stations located on unstable, heavily
subsiding ground and include stations on sturdy bedrock.
In building our time-dependent strain model, we first filter these
horizontal monthly median displacements using a technique
called “GPS Imaging.” This technique uses a median-spatial filter
to remove outliers, enhance the signal common to multiple
stations, and improve the overall robustness of the input
displacements for the strain model. An important property of this
technique is its ability to despeckle the displacement field while
resolving sharp edges between contrasting spatial domains. We
then use these filtered displacements to model our dilatational
and shear strain field for each month of the time series.
We validate each step of our procedure by examining the
potential pitfalls of our method. We test the resolution of the GPS
Imaging output by performing checkerboard reconstruction tests
to understand the geographic variation in spatial resolution of the
GPS measurement network. We additionally explore how varying
the a priori faulting constraints in the strain-rate tensor affects our
strain model.
Systematic fluctuations in the global seismic moment
release, Corné W Kreemer, and Ilya Zaliapin (Poster
Presentation 306)
We revisit the significance of the increased number of great
earthquakes since 2005. Analysis of the global moment release
during 1918–2012 shows that neither tapered nor doubly
truncated Pareto model with time-independent parameters can
satisfactorily explain the observations, suggesting decreased
moment release during 1975–2004. A model-independent
analysis, by resampling the observed moments, also indicates a
lower level of the moment release during this period. Our results
suggest the existence of temporal variation in the seismic
moment distribution on a global scale. The post-2004
observations signify a transition to a regime with increased
moment release. We argue that our conclusions remain valid for
alternative choices of moment distribution and its
parameterization, as long as a model is consistent with available
data. The analysis uses sequential order of events but not their
occurrence times, and hence the results are independent of a
particular model for the earthquake occurrence times.
Mitigating Induced Seismicity Through Active Pressure
Management: Simulation-Based Studies, Kayla A Kroll,
Keith B Richards-Dinger, and Joshua A White (Poster
Presentation 317)
The recent upturn of seismicity rates in the Central and Eastern
United States and Canada has been attributed to industrial
operations such as waste-water disposal, hydraulic fracturing,
and subsurface carbon storage. We couple the 3D, physics-
based earthquake simulator, RSQSim, to a reservoir model to
investigate the space-time characteristics of earthquakes
induced by pore-fluid pressure increases and/or poroelastic
stresses during injection. RSQSim employs rate-state friction,
which gives rise to spatiotemporal earthquake clustering. The
simulator generates long catalogs of seismicity based on stress
changes due to fault interaction and external stress perturbations
with great computational efficiency, allowing for multiple
simulations to systematically explore the parameters that control
induced seismicity. These simulations provide physics-based
statistical data that may contribute to the formalization of optimal
injection operations designed to minimize risk of seismicity at a
given industrial site.
Industrial operators may modify injection rates as an active
seismicity mitigation tool to either reduce the total number of
earthquakes or attempt to reduce the likelihood of future large
events. To explore the efficacy of this approach, we use RSQSim
to explore how sequences of induced earthquakes respond to
changes in injection schedule. We simulate induced seismicity
on a single optimally oriented fault with fractally distributed initial
shear stresses and compare results from models with/without
along-strike fault permeability and poroelastic stress changes.
We investigate the seismic response to several injection
schedules that lie between two end-member scenarios, 1)
constant injection at low rates, and 2) periodic injection at high
rates. We evaluate the cumulative number of events, total
seismic moment release, and the spatio-temporal characteristics
WELCOME
2016 SCEC Annual Meeting page 198
of seismicity including the time/location of the next large event
after adjusting injection rates.
Prepared by LLNL under Contract DE-AC52-07NA27344.
A Simulation Approach for Better Understanding of
Seismic Hazard and Risk in Montreal, Yelena
Kropivnitskaya, Kristy F Tiampo, Jinhui Qin, and Michael Bauer
(Poster Presentation 196)
Despite the fact that seismic activity occurs primarily along plate
boundaries, large earthquakes also occur within the plates and
result in significant human and economic losses (Adams and
Basham, 1989). Unlike seismicity along the plate boundaries,
intraplate seismicity is unevenly distributed and often poorly
understood. For example, the Charlevoix seismic zone, the most
seismically active region in eastern Canada is located in a stable
continental region within the North American Plate. While it has
a relatively low rate of earthquake activity, this largely urban
region has experienced six earthquakes of approximately
magnitude 6 since 1663, with large earthquakes concentrated in
regions of crustal weakness (GSC, 2016). Sykes (1978) found
that, in general, most events within northeastern America occur
in those areas affected by the last major orogenesis, before the
opening of the Atlantic Ocean. Eastern Canadian seismicity may
be the result of reactivation of existing fault zones and other
tectonic boundaries, such as in the late Proterozoic to Paleozoic
fault system along the St. Lawrence and Ottawa rivers (Thomas,
2006).
Montreal, located in the seismically active region of the western
Quebec seismic zone, is one the largest and most populated
cities in the eastern Canada. Here we present the steps and
results of a high-performance computing application to
probabilistic seismic hazard analysis (PSHA) for Montreal. The
simulations are performed based on the pipelining
implementation of the EqHaz program suite (Assatourians &
Atkinson, 2013) in the IBM InfoSphere Streams (Kropivnitskaya
et al., 2015). This pipelined hazard mapping approach identified
two main processing operators as bottlenecks: the catalog
generation operator using Monte Carlo simulations and the map
generation operator. The workload of the hazard mapping
workflow has been decomposed and these two operators have
been split into multiple pipelines for parallel execution. This
allows for the computation of hazard maps for different frequency
bands at significantly increased resolution with much lower
latency time for Montreal city. Moreover, a range of high-
resolution sensitivity tests are performed and conclude with
insights as to which input factors and parameters significantly
affect the ground motion results and are the main sources of
uncertainty. In addition, the potential application of streaming
techniques (Kropivnitskaya et al., 2015) to near-real time high-
resolution shakemap production for scenario or actual events is
demonstrated here. These results can be used by policymakers
to set earthquake resistant construction standards, by insurance
companies to set insurance rates and by civil engineers to
estimate stability and damage potential for Montreal city.
Seasonal motions and interseismic strain measured by
continuous GPS throughout the Transverse Ranges, CA,
Hannah E Krueger, Scott T Marshall, Susan E Owen, Gareth J
Funning, and Jack P Loveless (Poster Presentation 167)
Tectonic continuous GPS velocity estimates are complicated by
quasiperiodic motions that arise due to seasonal variations in
atmospheric conditions and/or anthropogenic activity. In
southern California, a dense GPS network, large seasonal
motions, and seasonally variable tropospheric conditions make
the region ideal for characterizing seasonal and anthropogenic
motions. In this study, we utilize data from 347 permanent GPS
sites in the Plate Boundary Observatory network. We calculate
periods and amplitudes of seasonal motion on detrended GPS
time series for each station. We find that seasonal ground
motions at stations near known groundwater extraction sites, oil
extraction, and other anthropogenic activity have significantly
different characteristics than the surrounding stations. This
suggests that identification of anomalous annual phases and/or
amplitudes in GPS time series may help to identify GPS stations
with signals contaminated by nontectonic sources. Based on this
analysis, we identify and discard numerous contaminated sites
from the data set and present an improved tectonic velocity
model for southern California. Then, to isolate deformation in the
Ventura and Los Angeles regions, we remove interseismic
deformation associated with the San Andreas, San Jacinto, and
Garlock faults using a dislocation model. The residual velocities
are then inverted for strain rates highlighting zones of
interseismic strain accumulation throughout the Transverse
Ranges region.
Exploring the sensitivity of multiscale seismic cycle
simulations to material heterogeneities., Christodoulos
Kyriakopoulos, Keith B Richards-Dinger, and James H Dieterich
(Poster Presentation 322)
We are presenting preliminary results on the estimation of the
effects of variable material properties in the earthquake simulator
RSQSim (Richards-Dinger and Dieterich, 2012). In order to do
that we are calculating elastostatic solutions (Green’s functions)
in a heterogeneous medium using the Finite Element (FE) code
Defmod (Ali, 2014) instead of the analytical solutions of Okada
(1992). The investigated region is a ~400 x 400 km area centered
on the Anza zone in southern California. The fault geometry is
extracted from the UCERF3 database, which we then implement
in a finite element mesh using Trelis 15. The structure of the
earth’s crust is based on available tomographic data for the
specific region. The Green’s functions are collected into the
“heterogeneous” stiffness matrix Khet and used in RSQSim
instead of the ones generated with Okada. Our sensitivity test is
based on the comparison between RSQSim results dependent
on the “heterogeneous” Khet and RSQSim results dependent on
Khom (FE model with homogeneous material properties). Our
preliminary simulations, although spatially centered in the San
Jacinto fault zone, will provide significant insights in order to, first,
quantify how much multiscale earthquake simulators are affected
by material heterogeneities, and second, design future
simulations based on FE methods and that will cover a larger
domain.
Evidence for strong lateral seismic velocity variation in the
lower crust – upper mantle beneath the California margin,
Voon Hui Lai, Robert W Graves, Shengji Wei, and Donald V
Helmberger (Poster Presentation 022)
Regional seismograms from earthquakes in Northern California
show a systematic difference in arrival times across central and
WELCOME
2016 SCEC Annual Meeting page 199
southern California where long period (30 – 50 seconds) SH
waves arrive up to 15 seconds earlier at stations near the coast
compared with sites towards the east at similar epicentral
distances. We attribute this time difference to heterogeneity of
the velocity structure at the crust − mantle interface beneath the
California margin. To explain these observations, we propose a
relatively simple 3D model consisting of a fast seismic lid with
thickness growing westward from the San Andreas fault, along
with a thicker and slower continental crust to the east. Synthetics
generated from such a model are able to better match the
observed timing of SH waveforms throughout most of central and
southern California compared with the 3D models currently
available in the SCEC UCVM. The presence of a strong lid
buttressed against a weaker crust has a major influence in how
the boundary between the Pacific plate and North American plate
deforms and is consistent with the observed asymmetric strain
rate across the boundary.
Rupture zone characteristics of the 2010 El Mayor-
Cucapah (Mexico) earthquake revealed with differential
lidar, Lia J Lajoie, and Ed Nissen (Poster Presentation 115)
Recent geodetic studies of surface deformation in large
earthquakes have started to make connections between rupture
characteristics and the mechanics, geometry, and material
properties of the host faulting. Here, we investigate these
relationships for the El Mayor Cucapah earthquake by integrating
a rich set of field measurements (Fletcher et al.,2014; Teran et
al., 2015) with a unique set of three-dimensional surface
displacements derived from pre- and post-event airborne lidar
topography (Glennie et al., 2014). First, in ground-rupturing
earthquakes, slip estimated at depths of 100s of meters to a few
kilometers typically exceeds the magnitude of offsets surveyed
along the primary fault trace (the “shallow slip deficit”). We seek
to clarify how fault zone structural maturity, width, geometry,
variations in near-surface material properties, or combinations of
these factors might affect this deficit. To establish the slip deficit
on each of an array of rupture segments, we compare offsets
surveyed at the principal fault scarp with lidar-derived surface
displacements measured short distances (100s of meters) from
the scarp, and outside of the damage zone observed in the field,
which are a proxy for slip at depth. Our second goal is to resolve
a discrepancy between field observations in the northern rupture
zone which implicate slip on low-angle, NE-dipping detachment
faults, and models based upon space geodetic and seismological
data which support only sub-vertical structures. Along most of the
fault trace, we find that the vertical throw accommodated across
the rupture zone far exceeds the extensional heave, indicating
that coseismic faulting at depth is steeply-dipping. However,
these relations are reversed along a ~5 km section of the Paso
Superior fault, where we our data support earlier field-based
inferences of low-angle slip. A third goal is to answer outstanding
questions about whether the rupture direction of an earthquake
is imprinted upon the pattern of surface deformation it produces.
Some theories suggest that faults may have a preferred rupture
direction that would preferentially concentrate deformation on the
same side of the fault for each earthquake. We address this by
projecting the mapped surface trace onto surface offset profiles
to assess damage zone asymmetry.
Ground Motion Simulation Validation using Small-to-
Moderate Magnitude Events in the Canterbury, New
Zealand Region, Robin L Lee, Brendon A Bradley, and Seokho
Jeong (Poster Presentation 286)
This poster presents preliminary results from ground motion
simulations of small-to-moderate magnitude (3<M_W<4.5)
earthquake events in the Canterbury region over the past
decade, for which centroid moment tensor solutions are
available. The simulations are carried out using SPECFEM3D,
an open source software package based upon the spectral-
element method, and both a regional 1D velocity model and the
recently developed 3D Canterbury Velocity Model (CantVM)
version 1. The simulated waveforms from both velocity models
are compared against observed waveforms at strong motion
stations; where it is seen that simulated ground motions utilizing
the 3D velocity model provide a better fit to the observed ground
motions than those from the 1D velocity model. The results of the
forward ground motion simulations presented are being
subsequently utilized in tomographic full waveform inversion to
improve the CantVM version 1.
Seafloor expression of active transpressional faulting
offshore Southern California, Mark R Legg, Simon L
Klemperer, Christopher M Castillo, Marie-Helene Cormier,
Michael Brennan, Katy Croff Bell, Dwight Coleman, Chris
Goldfinger, and Jason Chaytor (Poster Presentation 113)
Recent observations using Remotely Operated Vehicles (ROV)
of the seafloor along active transpressional fault zones offshore
southern California reveal the morphology of active fault ruptures
in the deep marine environment. Pressure ridges were first
identified using high-resolution multibeam bathymetry and
subsequently investigated using high resolution MCS surveys
conducted by Stanford and Oregon State university scientists.
These MCS data were used to guide ROV Hercules operated
from the Ocean Exploration Trust Exploration Vessel Nautilus.
These pressure ridges were located within long, narrow, hillside
valleys on the flanks of major transpressional uplifts including the
Santa Cruz-Catalina Ridge in the Inner Borderland and
Southwest Bank in the Outer Borderland. In general, the
pressure ridge is expressed in the bathymetry as an elongate and
somewhat sinuous ridge with relatively minor seafloor relief.
Although located within the sedimentary fill of significant hillside
valleys, these ridges are covered with angular blocks of bedrock,
inferred to be brecciated and squeezed upward during large
earthquake ruptures along the fault. In a sense, these may
represent the long term expression of mole tracks, common to
large strike-slip earthquake ruptures in alluvial basins onshore,
but better preserved over multiple earthquake cycles in the deep
marine environment where erosion is subdued and
sedimentation is slow. Along the flanks of the hillside valley are
steep to vertical scarps in bedrock outcrops between sediment
cover with angular blocks in talus slopes below. Scarps exposed
near the crest of large restraining bend pop-up structures show
complex fracturing and fault trends consistent with dextral shear
in the hanging wall of a moderate-to steep-dipping primary fault.
These seafloor observations are consistent with large-scale
transpressional structure identified in geophysical data (seismic
reflection surveys). The lateral scale of the features, >10 km, and
vertical relief 10-1000 meters (or greater) may represent
infrequent large earthquakes (M>7) and potential for local
tsunami generation. As yet, fault slip rates are unknown but the
WELCOME
2016 SCEC Annual Meeting page 200
new data may provide first-order estimates of deformation rates
for these prominent offshore structures.
Structural Architecture of the Western Transverse
Ranges and Potential for Large Earthquakes, Yuval Levy,
and Thomas K Rockwell (Poster Presentation 082)
Fold-and- thrust belts evolve over time, can produce large-scale
faults and potentially accommodate large magnitude
earthquakes. The thrust fronts of these structures typically form
large fold structures in their hanging walls, and they tend to
propagate forward over time to form new thrust fronts. In the
Santa Barbara and Ventura region of the Western Transverse
Ranges (WTR) of southern California, the Pitas Point Thrust is
interpreted as the current thrust front structure, and spatially
stable back thrusts accommodate deformation in the hanging
wall block of the thrust sheet (More Ranch fault, Rincon Creek
fault, other faults). In our work we constructed several cross-
sections for the WTR, combining various sources of data and
previous models suggested by others. We interpret the nearly
continuous overturned Tertiary stratigraphy of the Santa Ynez
Mountains as a large anticlinorium that formed as the first thrust
front over the (mostly) blind San Cayetano thrust, and that the
thrust front propagated south with time to the Red Mountain fault
and eventually to the currently active thrust front, the southward-
vergent Pitas Point-Ventura fault. We further suggest that the
steep dip angle of the Red Mountain fault, as observed near the
surface, is a result of northward rotation of the fault, which causes
it to flexurally slip. The northward rotation is also responsible for
continued folding to the north (Ayers Creek syncline) and back
thrusting on the hanging wall of the Red Mountain fault (Arroyo
Parida fault).
Tremor and LFE activities using mini seismic array in the
Alaska-Aleutian subduction zone, Bo LI, and Abhijit Ghosh
(Poster Presentation 215)
Tremors and low frequency earthquakes (LFEs) have been well
studied and observed as coupled phenomena in several
subduction zones, such as the southwestern Japan, Cascadia
and Costa Rica [Ghosh et al., 2012; Shelly et al., 2006; Brown et
al., 2009]. In the Alaska-Aleutian subduction zone, the harsh
weather conditions, the limited land and seismic station
coverage, and the interference of the volcanic activities make it
difficult to study tectonic (non-volcanic) tremor and LFE activities.
In this study, we use continuous seismic data recorded by a mini
array for two months to detect and locate tremors and LFEs. The
array is composed of 11 stations with three components,
deployed on Akutan Island in 2012. We use the beamprojection
method [Ghosh et al., 2009; Ghosh et al., 2012] to automatically
scan the continuous array data for tremors, and detect an
average of 1.3 hours tremor activities per day. Then we use the
local earthquakes recorded in the Advanced National Seismic
System (ANSS) Composite Earthquake catalog during this
period in the study region to calibrate the array. After calibration,
the tremors locate into two clusters. The major one is located at
the southern edge of the Unalaska Island and a smaller cluster
to the southeast of the Akutan Island, with a gap around 30 km
between them. Besides several visually detected LFEs, we apply
a method similar to the beamforming network response method
[Frank et al., 2014] to detect the LFEs using a sliding time window
of 15 s, with a 0.025 s step. We detect hundreds of LFEs and
classify them into 11 families. Then we use the LFEs detected as
templates to search for the repeating LFEs over the two-month
period by applying the matched-filter method. During the tremor
activity, the LFEs show a much shorter interval time than non-
tremor period. The locations of LFEs using first arrivals of P- and
S-waves match well with the tremor clusters. To better
understand the subduction fault dynamics, an additional mini
array was installed in 2014 and another two are deployed in the
summer 2015. The improved azimuth coverage enables us to
detect and locate more tremors and LFEs, and provides us a
good opportunity to study the spatiotemporal distribution of slow
and fast earthquakes, and their potential connection in the
Alaska-Aleutian subduction zone.
Improved Understanding of Triggered Seismicity in the
Salton Sea Geothermal Field with Waveform Matching
Method, Chenyu Li, Zhigang Peng, Dongdong Yao, Bridget
Casey, and Xiaofeng Meng (Poster Presentation 214)
Microseismicity can be easily triggered in volcanic and
geothermal regions by stress perturbations from earthquake
waves hundreds to thousands of kilometers away. Geothermal
fields are seismically active with intensive microseismicity, and
highly sensitive to external stress perturbation, making them
ideal “natural laboratory” for studying earthquake triggering. The
Salton Sea Geothermal Field (SSGF) is one of the most
seismically active and geothermally productive fields in
California. It lies close to major active faults such as the Southern
San Andreas Fault, Imperial Fault, Brawley Seismic Zone (BSZ)
and San Jacinto Fault. Previous studies already found evidence
that seismicity in SSGF was triggered by 1999 M7.1 Hector Mine
Earthquake and 2010 El Mayor-Cucapah Earthquake. Here we
present a systematic investigation of triggered seismicity in
SSGF from 2007 to 2014 utilizing the Calenergy Borehole
Network (EN). We apply the recently-developed GPU-based
waveform matched-filter technique (WMFT) to detect missing
microearthquakes and analyze the seismicity change in SSGF
around some regional and remote earthquakes with predicted
dynamic stress above 5 KPa. Our result shows triggered
seismicity following a few regional earthquakes, such as the 2009
Mw6.9 Baja California and 2010 Mw5.7 Ocotillo Earthquakes,
and it tends to be less sensitive to remote large earthquakes
(e.g., 2010 Mw8.8 Chile and 2011 Mw9.1 Tohoku earthquakes).
Our next step is to apply the same procedures to mainshocks
that are even closer, such as the 2012 BSZ earthquake swarms
that are ~20 km south of SSGF, to examine whether the SSGF
respond to stress perturbations from nearby earthquakes.
Updated results will be presented in the meeting.
Fault Damage Zone of the 2014 Mw 6.0 South Napa
Earthquake, California, Viewed from Fault-Zone Trapped
Waves, Yong-Gang Li, Rufus D Catchings, and Mark R
Goldman (Poster Presentation 176)
Prominent fault-zone trapped waves (FZTWs) were generated by
aftershocks of the 2014 M6 South Napa earthquake and
recorded at three dense linear arrays across the surface rupture
of the West Napa Fault Zone (WNFZ) and the Franklin fault,
which appears to be southward continuations of the WNFZ. We
analyzed waveforms of FZTWs from 55 aftershocks in both time
and frequency to characterize the fault damage zone associated
with this M6 earthquake. Post-S coda durations of FZTWs
WELCOME
2016 SCEC Annual Meeting page 201
increase with epicentral distances and focal depths from the
recording arrays, suggesting a low-velocity waveguide along the
WNFZ to depths in excess of 5-7 km. Locations of the
aftershocks showing FZTWs, combined with three-dimensional
(3D) finite-difference simulations, suggest the subsurface rupture
zone having a S-wave speed reduction of ~40-50% and at least
a ~500-m-wide deformation zone along with the ~14-km-long
mapped surface rupture zone. The low-velocity waveguide on the
WNFZ extends further southward but with a more moderate
velocity reduction of ~30% at ray depth, suggesting continuity
between the WNFZ and Franklin Fault. Recently, we examined
the data recorded at 7 REFTEK130 seismometers deployed near
the WNFZ and obtained additional constraints on the results from
the data recorded at three cross-fault arrays, implying that the
waveguide effect may have localized and amplified ground
shaking along the WNFZ and the faults farther to the south. We
try to search the FZTWs recorded at these temporal stations at
the WNFZ for teleseismic earthquakes. Such type of FZTWs has
been observed at the seismic array deployed atop the Calico
Fault in Mojave Desert for teleseismic events; they provide an
additional constraint on the deep portion of the fault compliant
zone at seismogenic depths in excess of ~8-km that was imaged
previously (Cochran et al., 2009).
Two decades of shear-wave splitting measurements in
southern California, Zefeng Li, and Zhigang Peng (Poster
Presentation 191)
We measure shear-wave splitting parameters (i.e. fast direction
and delay time) of local earthquakes recorded by the Southern
California Seismic Network from 1995 through 2014. A set of
codes is used to automatically download three-component
waveform data from the Southern California Earthquake Data
Center (SCEDC), pick S-wave arrivals, and measure the splitting
parameters without any human interference. We also calculate
signal-to-noise ratio and measurement uncertainties and use
them as quality control parameters. Initial check on some stations
that were frequently studied (e.g. AZ.KNW) shows very good
agreement between the new results and the previous ones using
relatively short period of data. We are in process of cleaning up
and compiling all the results. After that, we plan to analyze
spatiotemporal variations of splitting parameters and compare
them with regional stress measurements and community fault
models to better understand physical mechanisms of crustal
anisotropy in southern California. Updated results will be
presented in the meeting.
Isochron burial dating of paleosols within the Whitewater
Fan, northern Coachella Valley, California, Nathaniel Lifton,
Richard V Heermance, Doug Yule, and Brittany Huerta (Poster
Presentation 126)
Remnant alluvial fan surfaces are ubiquitous features across the
San Gorgonio Pass (SGP), distinguished by their perched
location 50-150 m above the active channels and dark red,
>1.5m thick soil horizons. Their potential utility in constraining
long-term slip rates along the southern San Andreas fault is
widely recognized, but their age is still debated. This study
focuses on one of these surfaces – the upper surface of the
Whitewater Fan (aka Whitewater Hill) – which has been uplifted
and deformed between the Banning and Garnet Hill strands of
the San Andreas Fault in the northern Coachella Valley, CA.
Mapping of the Whitewater Hill area in summer 2015 revealed at
least 2 buried paleosols (>70 m below the surficial red soil)
defining a long-term record of alluvial fan development. These
paleosols can be used as piercing points to develop long-term
slip rates along this section of the San Andreas fault system if
their ages can be determined. We collected a total of 18 samples
from both buried paleosols for cosmogenic 10Be-26Al isochron
burial dating. This technique derives its power from sampling
multiple cobbles from a narrow depth range, each of which has
experienced the same postburial history. The presence of the
paleosols (albeit truncated) indicates unambiguous subaerial
exposure and subsequent burial. If each cobble has a different
exposure history prior to burial, then the samples will define a line
on a plot of 26Al relative to 10Be, with a slope equal to the ratio
of the two nuclides at production. Burial shuts off surficial nuclide
production and leads to decay of 26Al (t1/2 ca. 0.7 My) relative
to 10Be (t1/2 ca. 1.4 My). This differential decay causes the slope
of the line to decrease in a predictable manner with increasing
burial age. Since all clasts have experienced the same postburial
history, this method is less sensitive to effects of surface deflation
than methods using near-surface clasts. Be and Al were
extracted at the Purdue Rare Isotope Measurement (PRIME) Lab
and analyzed by Accelerator Mass Spectrometry (AMS). A
paleosol exposed in the east wall of Whitewater Wash yielded 11
cobbles with enough quartz for analysis, while only 3 of 7 cobbles
collected from a paleosol exposed in a small drainage to the west
of the main wash yielded enough quartz of high enough purity
(low enough total Al concentration) for analyses. Results suggest
systematic complexities that are currently being evaluated to
shed light on fan depositional processes and ages.
Topographic Influence on Near-Surface Seismic Velocity
in southern California, Jessica C Lin, Seulgi Moon, Lingsen
Meng, and Paul M Davis (Poster Presentation 177)
Near-surface seismic velocity is commonly used to determine
subsurface rock structure, properties, and ground-motion
amplification. The spatial distribution of Vs30 (shear-wave
seismic velocity in the top 30 m of Earth’s crust) has been inferred
based on the correlations of measured Vs30 with rock types and
topographic slopes. Inference of Vs30 based on topographic
slopes relies on the assumption that mechanically strong rocks
tend to have steep slopes. The topographic slopes can thus be
used to infer bedrock strength and seismic velocity. However,
due to limited accessibility and logistical difficulties, there are few
Vs30 measurements in sites of crystalline rocks that have
measurable topographic variations. Thus, the variability of Vs30
with topographic slope for crystalline rocks has not been
addressed systematically. In order to examine the local
variabilities in near-surface seismic velocity in southern
California, we measured the spatial distributions of near-surface
seismic velocity at two sites: one in the San Gabriel Mountains
(SGM) and one in the San Bernardino Mountains (SBM). Both
sites are composed of predominantly crystalline rocks with
topographic slopes that range from 0.2 to 0.5. We conducted
seismic refraction surveys using sledgehammer-induced impacts
on a steel plate along seismic lines that were oriented roughly N-
S, 240 m in length with a spacing of 5 m, and with topographic
variation including both a local hilltop and valley. Using first P-
wave arrivals, we constructed a P-wave seismic tomography
down to 50 m. Our results show that P-wave seismic velocity in
WELCOME
2016 SCEC Annual Meeting page 202
the SGM site varies significantly within hillslopes and does not
linearly correlate with slope, while P-wave seismic velocity in the
SBM site shows little variation in the hillslope. In the SGM site,
the Vs30 beneath the valley is 25% faster than the Vs30 beneath
the hillslope. These results suggest that the local variability of
seismic velocity depends on differences in sediment thickness,
bedrock fractures, and weathering patterns.
Evidence for Non-Self-Similarity of Microearthquakes
Recorded at a Taiwan Borehole Seismometer Array, Yen-
Yu Lin, Kuo-Fong Ma, Hiroo Kanamori, Teh-Ru A Song, Nadia
Lapusta, and Victor C Tsai (Poster Presentation 212)
We investigate the relationship between seismic moment M0 and
source duration tw of microearthquakes by using high-quality
seismic data recorded with a vertical borehole array installed in
central Taiwan. We apply a waveform cross-correlation method
to the three-component records and identify several event
clusters with high waveform similarity, with event magnitudes
ranging from 0.3 to 2.0. Three clusters—Clusters A, B, and C—
contain 11, 8, and 6 events with similar waveforms, respectively.
To determine how M0 scales with tw, we remove path effects by
using a path-averaged Q. The results indicate a nearly constant
tw for events within each cluster, regardless of M0, with mean
values of tw being 0.058, 0.056, and 0.034 s for Clusters A, B,
and C, respectively. Constant tw, independent of M0, violates the
commonly used scaling relation . This constant duration may
arise either because all events in a cluster are hosted on the
same isolated seismogenic patch, or because the events are
driven by external factors of constant duration, such as fluid
injections into the fault zone. It may also be related to the
earthquake nucleation size.
Apparent Attenuation at High Frequencies in Southern
California, Yu-Pin Lin, and Thomas H Jordan (Poster
Presentation 228)
Accurately simulating strong motions for seismic hazard analysis
requires accurate 3D models of crustal structure. At low
frequencies (< 1 Hz), the amplitude reduction of seismic
wavefields due to anelastic attenuation is relatively minor, and
available tomographic models for Southern California (e.g.,
CVM-S4.26 of Lee et al., 2014) do a pretty good job of accounting
3D elastic scattering on wavefield amplitudes. At higher
frequencies, however, anelastic attenuation becomes more
important, and the elastic scattering depends on unresolved
small-scale heterogeneities, giving rise to a complex apparent
attenuation structure that depends on both position and
frequency. We place constraints on this structure in the band 1-
10 Hz through the analysis of earthquake waveforms recorded
by the Southern California Seismic Network (SCSN). We localize
signals in frequency and time using wavelet transforms, and we
account for source structure and geometrical spreading by
referencing the spectral amplitudes to values computed from
synthetic seismograms. This approach is examined by the
synthetic test which is accomplished by F-K method. We directly
invert narrow-band spectral amplitudes rather than the t*
measurements commonly used in previous studies, which allows
better resolution of frequency-dependent effects. Inversions of
datasets recover an attenuation structure that, when averaged
laterally and over frequency, is generally consistent with the
tomographic study of Hauksson & Shearer (2006). In particular,
we find that the apparent quality factor for P waves (QP) is less
than the apparent quality factor for S waves (QS), in contradiction
with the classical relation QP ~ 2QS that has been used for most
wavefield modeling at low frequencies. The data are consistent
with QP anomalies being strongest in the low-Q, near-surface
waveguide, suggesting that strong scattering from small-scale
heterogeneities may play a role in explaining this discrepancy.
The data also require that the apparent attenuation be strongly
frequency dependent across the 1-10 Hz band.
Mapping fault creep, frictional properties, and
unrecognized active structures with dense geodetic data
in the Imperial Valley, Southern California, Eric O Lindsey,
and Yuri Fialko (Poster Presentation 129)
The Imperial-Mexicali valley is well known as an area of
significant earthquake hazard, but it remains a challenging target
for geodetic studies because of agricultural and geothermal
activity which obscures both short- and long-term deformation
and can lead to aquifer-related contamination of the geodetic
signals. We address these difficulties by combining InSAR data
from four independent Envisat tracks, processed using a
persistent scatterers method, to separate the vertical and
horizontal (fault-parallel) motion. We combine these data with
coseismic and postseismic geodetic observations of the 1979 M
6.6 Imperial Valley earthquake (Harsh, 1982 and Crook et al.,
1982) and dense interseismic GPS velocities (Crowell et al.,
2013). The result is a dense map of geodetic observations
spanning nearly a complete earthquake cycle on the Imperial
fault, which allows us to evaluate the rate of interseismic loading
and along-strike variations in surface creep. We compare the
data to dynamic models with rate- and state-dependent friction
and show that a complete record of the earthquake cycle is
required to constrain key fault properties including the rate-
dependence parameter (a − b) as a function of depth, the extent
of shallow creep, and the recurrence interval of large events.
In addition, we show that the data are inconsistent with a high
(>30 mm/yr) slip rate on the Imperial Fault and investigate the
possibility that an extension of the San Jacinto-Superstition Hills
Fault system through the town of El Centro may accommodate a
significant portion of the slip previously attributed to the Imperial
Fault. Such a fault has previously been proposed (Thomas &
Rockwell, 1996 and Hogan et al., SCEC abstract, 2002).
Dislocation models including this additional fault are in better
agreement with the available observations, suggesting that the
long-term slip rate of the Imperial Fault is lower than previously
suggested and that there may be a significant unmapped hazard
in the western Imperial Valley.
Tremor and Slow Slip in the West Antarctic Ice Sheet,
Bradley P Lipovsky, and Eric M Dunham (Poster Presentation
072)
The Whillans Ice Plain region of the West Antarctic Ice Sheet
experiences twice-daily, tidally-modulated slow slip events.
During each event, 0.5 m of slip occurs over a 150x150 km area.
Sliding initiates at one of several recurring locations and expands
outwards with a typical rupture velocity ~200 m/s. In addition,
slow slip events are accompanied by seismic tremor episodes.
Understanding the mechanical processes that gives rise to these
curious phenomena has implications for both sea level rise
WELCOME
2016 SCEC Annual Meeting page 203
hazard and for our knowledge of faulting processes more
generally.
We model these observations using a simplified model that
includes the effects of inertia, elasticity, tidal forcing, and rate-
and state-dependent frictional sliding. We interpret seismic
tremor during slow slip events as due to the failure of m-scale
rate weakening patches at the ice-bed interface. We find that, at
the ~km-scale, slow rupture velocities are explained by basal
conditions that are near the steady sliding limit. Such conditions
arise in our simulations due to the presence of high pore
pressures such that the subglacial effective normal stress is ~10
kPa where the ice overburden pressure is 7.2 MPa.
Constraining the velocity structure near the tremorogenic
portion of the San Andreas Fault near Parkfield, CA using
Ambient Noise Tomography, Rachel C Lippoldt, Robert W
Porritt, and Charles G Sammis (Poster Presentation 253)
The central section of the San Andreas Fault (SAF) displays a
range of seismic phenomena including tectonic tremor, low
frequency earthquakes (LFE), repeating micro-earthquakes
(REQ), and aseismic creep. We investigate the seismic velocity
structure associated with this section of the SAF using surface
wave tomography from ambient seismic noise. A high-resolution
model of shear velocities in the crust and upper mantle is derived
from data recorded between 2003 and 2015 by the USArray, the
Parkfield Hi-Resolution Seismic Network (HRSN), and other
regional networks. Fault perpendicular sections show that the
SAF is revealed by lateral contrasts in relative velocities to
depths to 50 km. At depths between about 15 and 30 km, the
deep extension of the SAF appears to be localized to within about
50 km. This localization is consistent with the hypothesis that
LFEs are shear-slip events on a deep extension of the SAF. Fault
parallel sections show that LFE and REQ locations occur within
low velocity structures, suggesting that the presence of fluids,
weaker minerals, or hydrous phase minerals may play an
important role in the generation of slow slip phenomena. Such
correlations between seismic waves speeds and seismicity can
provide insight and constraints on the physical mechanics that
produce LFEs and REQs.
Stress heterogeneity at restraining double bends under
multicycles and its effect on rupture propagation in 3D,
Dunyu Liu, and Benchun Duan (Poster Presentation 050)
The conditions under which a rupture is able to propagate
through geometrical complexity, such as restraining double
bends, are crucial to the estimation of potential maximal event
size on a fault system like the central Altyn Tagh Fault. Previous
studies using 2D multicycle dynamic rupture models, including a
viscoelastic model to represent off-fault deformation and inter-
seismic loading and a spontaneous dynamic rupture for
coseismic deformation, illustrate that heterogeneous stress
builds up around restraining double bends under multiple
earthquake cycles. The restraining double-bend stops or initiates
rupturing in most cases but sometimes it may allow the rupture
link through the bends, leading to a large event. With the slip-
weakening friction law and uniform pre-stresses applied, stress
heterogeneity concentrates around the double-bend and a stable
rupture pattern develops after a number of cycles. However, the
2D models are assumed at a specific depth, at which depth the
rupture is limited, whereas the depth dependency of stress states
and of frictional property on fault may play a role in affecting the
rupture propagation. In this study, we extend the 2D methodology
to 3D case and apply it to the central Altyn Tagh Fault, whose
trace is geodetically measured. Preliminary results show that
ruptures tend to link through a double-bend at shallow depths if
a constant slip-weakening distance were applied over the whole
fault. However, frictional strengthening behavior at shallow fault
significantly affects the jumping condition of ruptures at the
double-bend and the consequent rupture behavior during
multiple earthquake cycles. Besides, the stress heterogeneity
observed over a large range of scales, is not well reproduced by
the 2D methodology with slip-weakening law and uniform pre-
stresses. Two ways have been proposed to be able to produce
the stress heterogeneity by dynamic rupture models. One is to
start from uniform pre-stresses and to include rate-weakening in
the friction law, whereas the other one keeps using the slip-
weakening law while directly replaces the uniform pre-stress to a
heterogeneous one. Both schemes are able to generate healing
phase during the rupture and resultant pulse-like rupture and
stress heterogeneity observed. In the current stage, we will apply
a self-similar pre-stress together with the simple slip-weakening
law to examine whether or not multiple scales of stress
heterogeneity can be reproduced and can be conserved during
multiple earthquake cycles, as well as effect of the stress
heterogeneity on rupture behavior around restraining double
bends.
Cumulative offsets revealed by airborne LiDAR along a
“creeping” section of the Haiyuan fault, northern Tibetan
plateau, Jing Liu, Tao Chen, Yanxiu Shao, Peizhen Zhang,
Kenneth W Hudnut, Qiyun Lei, and Zhanfei Li (Poster
Presentation 080)
Airborne LiDAR is a powerful and efficient tool for acquiring
topographic data with cm- to mm-resolution, a resolution high
enough for detailed description of landforms and their small
displacements along faults. It is therefore very appealing and
increasingly applied to active fault studies worldwide. However,
such data is largely lacking in China despite that there are many
active faults in the country. We recently acquired ~130 km of
airborne LiDAR topographic data in 1 km-wide swath along the
Haiyuan fault. LiDAR point cloud obtained has a density of 4-5
pts/m2 on average, sufficient to build 0.5-1 meter DEMs, reveals
the Haiyuan fault with unprecedented clarity. The Haiyuan fault
is one of the major continental left-lateral faults in northern
Tibetan plateau, with repeated large-magnitude earthquakes, the
most recent large one being the great 1920 M~8 Haiyuan
earthquake with a 230-km-long surface rupture (Zhang et al.,
1987). However, the 30 km-long section of fault beyond the
western termination of the 1920 rupture behaves differently, with
a combination of different modes of slip, from creeping (Cavalier
et al., 2008; Jolivet et al., 2014), intense microseismicity, to
possibly 1888 Jintai earthquake with estimated magnitude of
63/4 - 7. LiDAR data illuminates fine features of the fault
geometry, which can be quantified in strike deviation and fault
zone width. Numerous offset gullies are delineated, yielding a
wide range of offsets, from 50 m to as low as 2 m. The grouping
of the smallest offsets, 2-3 m, is the most obvious, despite its
small magnitude, lends support to the inference of historical 1888
Jintai earthquake rupturing to the surface (Zhou et al., 1992).
WELCOME
2016 SCEC Annual Meeting page 204
Comparing to the Parkfield section of the San Andreas fault, the
creeping section of the Haiyuan fault preserves more discrete
cumulative offsets, which deserves further discussion in terms of
long-term geomorphic imprints of creep versus stick-slip faulting
styles.
Seismic Risk from Induced and Natural Earthquakes for
the Central and Eastern United States, Taojun Liu, Nicolas
Luco, and Abbie B Liel (Poster Presentation 172)
Earthquake rates in some parts of the Central and Eastern United
States have increased dramatically in the past few years, mostly
due to wastewater injection associated with oil and gas activities.
This induced seismicity has caused damage to buildings and
raised substantial public concern. The U.S. Geological Survey
(USGS) published a seismic hazard model in March 2016 that
accounts for this elevated seismicity. The model provides a one-
year forecast for seismic hazard from both induced and natural
earthquakes. However, the implications of the hazard model for
seismic risk is unknown. This paper quantifies seismic risk due
to induced seismicity by combining the USGS 2016 model and
fragility curves representing collapse risk and risk of
nonstructural falling hazards for code-designed structures. This
assessment shows substantial increase in risk for some areas
compared to that due to natural earthquakes alone. The increase
in risk depends on building period, location, the performance
target of interest, and hazard modeling assumptions. For
exploratory purposes only, we also investigate the increase in
design ground motion levels that would correspond to the
increased risk.
Estimating amplitude uncertainty for normalized ambient
seismic noise cross-correlation with examples from
southern California, Xin Liu, Gregory C Beroza, and Yehuda
Ben-Zion (Poster Presentation 188)
We estimate the frequency-dependent amplitude error of
ambient noise cross-correlations based on the method of Liu et
al. (2016) for different normalizations. We compute the stacked
cross spectrum of noise recorded at station pairs in southern
California by averaging the cross spectrum of evenly spaced
windows of the same length, but offset in time. Windows with
signals (e.g. earthquakes) contaminating the ambient seismic
noise are removed as statistical outliers. Standard errors of the
real and imaginary parts of the stacked cross-spectrum are
estimated assuming each window is independent. The
autocorrelation of the sequence of cross-spectrum values at a
given frequency obtained from different windows are used to test
the independence of cross-spectrum values in neighboring time
windows. For frequencies below 0.2 Hz, we find temporal
correlation in the noise data. We account for temporal correlation
in computation of errors using a block bootstrap resampling
method. The stacked cross-spectrum and associated amplitude
are computed under different normalization methods including
deconvolution and whitening applied before or after ensemble
average of cross-spectrum values. We estimate the amplitude
errors based on error propagation from errors of stacked cross-
spectrum and verified by bootstrap method. We propose to use
this characterization of amplitude uncertainty to constrain
uncertainties in ground motion predictions based on ambient-
field observations.
Combination of GPS and InSAR Data for Crustal
Deformation Mapping, Zhen Liu, Zheng-Kang Shen, and
Cunren Liang (Poster Presentation 140)
We are developing an approach to integrate GPS and InSAR
data to generate 3-dimensional crustal motion map. Point-based
discrete GPS measurements are first interpolated to produce
continuous 3-D vector map at common grids covered by the
InSAR data, based on an algorithm of Shen et al. [2015] that
takes into account GPS station distance, network density and
configuration for data weighting. A Gaussian distance weighting
function and a Voronoi cell spatial weighting function are used in
the interpolation, the amount of weighting and degree of
smoothing can be spatially variable and optimally determined
based on in situ data strength. At the locations where both InSAR
and interpolated GPS data are available, we solve for 3-D
velocity field using a weighted least square method. For the
interpolated GPS data, we re-estimate uncertainties by adopting
a uniform smoothing to ensure that the uncertainty estimates are
not biased by uneven degree of smoothing. We investigate the
effects of two adjustable parameters (total weight of the site
weighting functions and spatial smoothing distance) on resultant
uncertainty estimates of the interpolated velocity field and test
the method using line-of-sight velocity from selected InSAR
tracks in southern California and a combination of continuous
and campaign GPS data. In addition, we present initial results of
imaging crustal deformation using new satellite sensor data such
as ALOS-2 ScanSAR. We show that it is crucial to correct
ionospheric noise source for L-band SAR data in order to get
accurate measurement of crustal motion induced by both tectonic
and non-tectonic sources.
Recent Achievements of the Collaboratory for the Study
of Earthquake Predictability, Maria Liukis, Maximilian J
Werner, Danijel Schorlemmer, John Yu, Philip J Maechling,
David D Jackson, David A Rhoades, Zechar D Zechar, Warner
Marzocchi, Thomas H Jordan, and the CSEP Working Group
(Poster Presentation 315)
The Collaboratory for the Study of Earthquake Predictability
(CSEP) supports a global program to conduct prospective
earthquake forecasting experiments. CSEP testing centers are
now operational in California, New Zealand, Japan, China, and
Europe with 442 models under evaluation. The California testing
center, started by SCEC, Sept 1, 2007, currently hosts 30-
minute, 1-day, 3-month, 1-year and 5-year forecasts, both alarm-
based and probabilistic, for California, the Western Pacific, and
worldwide. Our tests are now based on the hypocentral locations
and magnitudes of cataloged earthquakes, but we plan to test
focal mechanisms, seismic hazard models, ground motion
forecasts, and finite rupture forecasts as well. We have increased
computational efficiency for high-resolution global experiments,
such as the evaluation of the Global Earthquake Activity Rate
(GEAR) model, introduced Bayesian ensemble models, and
implemented support for non-Poissonian simulation-based
forecasts models. We are currently developing formats and
procedures to evaluate externally hosted forecasts and
predictions. CSEP supports the USGS program in operational
earthquake forecasting and a DHS project to register and test
external forecast procedures from experts outside seismology.
We found that earthquakes as small as magnitude 2.5 provide
important information on subsequent earthquakes larger than
WELCOME
2016 SCEC Annual Meeting page 205
magnitude 5. A retrospective experiment for the 2010-2012
Canterbury earthquake sequence showed that some physics-
based and hybrid models outperform catalog-based (e.g., ETAS)
models. This experiment also demonstrates the ability of the
CSEP infrastructure to support retrospective forecast testing.
Current CSEP development activities include adoption of the
Comprehensive Earthquake Catalog (ComCat) as an authorized
data source, retrospective testing of simulation-based forecasts,
and support for additive ensemble methods. We describe the
open-source CSEP software that is available to researchers as
they develop their forecast models. We also discuss how CSEP
procedures are being adapted to intensity and ground motion
prediction experiments as well as hazard model testing.
Evaluating the use of declustering for induced seismicity
hazard assessment, Andrea L Llenos, and Andrew J Michael
(Poster Presentation 309)
The recent dramatic seismicity rate increase in the central and
eastern US (CEUS) has motivated the development of seismic
hazard assessments for induced seismicity (e.g., Petersen et al.,
2016). Standard probabilistic seismic hazard assessment
(PSHA) relies fundamentally on the assumption that seismicity is
Poissonian (Cornell, BSSA, 1968); therefore, the earthquake
catalogs used in PSHA are typically declustered (e.g., Petersen
et al., 2014) even though this may remove earthquakes that may
cause damage or concern (Petersen et al., 2015; 2016). In some
induced earthquake sequences in the CEUS, the standard
declustering can remove up to 90% of the sequence, reducing
the estimated seismicity rate by a factor of 10 compared to
estimates from the complete catalog. In tectonic regions the
reduction is often only about a factor of 2. We investigate how
three declustering methods treat induced seismicity: the window-
based Gardner-Knopoff (GK) algorithm, often used for PSHA
(Gardner and Knopoff, BSSA, 1974); the link-based Reasenberg
algorithm (Reasenberg, JGR,1985); and a stochastic
declustering method based on a space-time Epidemic-Type
Aftershock Sequence model (Ogata, JASA, 1988; Zhuang et al.,
JASA, 2002). We apply these methods to three catalogs that
likely contain some induced seismicity. For the Guy-Greenbrier,
AR earthquake swarm from 2010-2013, declustering reduces the
seismicity rate by factors of 6-14, depending on the algorithm. In
northern Oklahoma and southern Kansas from 2010-2015, the
reduction varies from factors of 1.5-20. In the Salton Trough of
southern California from 1975-2013, the rate is reduced by
factors of 3-20. Stochastic declustering tends to remove the most
events, followed by the GK method, while the Reasenberg
method removes the fewest. Given that declustering and choice
of algorithm have such a large impact on the resulting seismicity
rate estimates, we suggest that more accurate hazard
assessments may be found using the complete catalog.
InSAR observations in Southern California, Rowena B
Lohman, and Kyle D Murray (Poster Presentation 137)
We present InSAR time series results over Southern California,
with a focus on the comparison between imagery from different
platforms and spanning different time periods in the Riverside
area. We use a processing approach that combines aspects of
several other SAR/InSAR/time series analysis packages,
including ROI_PAC, ISCE, StaMPS and phase unwrapping with
SNAPHU. This approach is not really necessary in the urban
areas of Los Angeles, but does appear to improve results in the
higher relief areas surrounding the city.
Dynamic Models of Large Ruptures on the Southern San
Andreas Fault, Julian C Lozos (Poster Presentation 061)
The southern San Andreas Fault is a heterogeneous structure.
Despite being a primary plate boundary fault, it has major
geometrical complexities in the Big Bend and San Gorgonio
Pass, and also shows substantial variation in maximum
horizontal stress orientation along strike. The earthquakes of
1857 and 1812 may have had at least one rupture endpoint at a
point of stress or geometrical complexity, and paleoseismic
records suggest that many prehistoric events also exhibit
geometrical segmentation. However, the possibility of a rupture
from Cholame to the Salton Sea is allowed by paleoseismic
records and multi-cycle models, and is considered in hazard
assessment. In this study, I use the 3D finite element method to
conduct models of dynamic rupture on the entire southern San
Andreas Fault. I implement stress orientations from seismicity
literature, but vary initial stress amplitudes in order to determine
which values are necessary to replicate historic or well-
documented paleoseismic ruptures, as well as which values and
nucleation locations are necessary to produce a wall-to-wall
rupture.
Nowcasting Oklahoma and The Geysers, California, Molly
Luginbuhl, John B Rundle, and Donald L Turcotte (Poster
Presentation 207)
Nowcasting is a new method of statistically classifying seismicity
and seismic risk (Rundle et al., 2016). In this paper, the method
is applied to the induced seismicity in Oklahoma and the induced
seismicity at The Geysers geothermal region in California.
Nowcasting utilizes the catalogs of seismicity in these regions.
Two earthquake magnitudes are selected, one small say M > 2,
and one large say M > 4. The method utilizes the number of small
earthquakes that occur between large earthquakes. The
cumulative probability distribution of these values is obtained.
The earthquake potential score (EPS) is defined by the number
of small earthquakes that have occurred since the last large
earthquake, the point where this number falls on the cumulative
probability distribution defines the EPS. A major advantage of
nowcasting is that it utilizes “natural time”, earthquake counts,
between events rather than clock time. For this reason, it is not
necessary to decluster aftershocks and the results are applicable
if the level of induced seismicity varies in time. The difference in
results for The Geysers, Oklahoma and examples of tectonic
earthquakes will be given.
Spontaneous Dynamic Rupture Simulation on
Geometrically Complex Faults Governed by Different
Friction Laws, Bin Luo, and Benchun Duan (Poster
Presentation 048)
The constitutive law describing how friction resistance evolves
dynamically on the existing shear fault surface is one of the major
factors controlling the behavior of dynamic earthquake rupture.
The laboratory-derived rate- and state- friction laws have been
widely applied in spontaneous dynamic rupture modeling and
behave similarly to the standard slip-weakening friction law on a
flat fault model. However, when geometrical complexity is
introduced to the fault surface, rupture behavior obeying different
WELCOME
2016 SCEC Annual Meeting page 206
friction laws may be affected differently by these geometrical
irregularities. Using our explicit, dynamic finite element method
code EQdyna, we carry out multiple spontaneous dynamic
rupture simulations on both planar and non-planar faults
governed by slip-weakening law(SW) and rate- and state- friction
laws including ageing law(RS-A), slip law(RS-S) and slip law with
flash heating(RS-FH) for comparison to explore the differences
between their associated rupture behaviors. Our result suggests
that models with all kinds of friction laws applied here share
common, friction-law-independent features of geometrical effects
on rupture behavior, such as local static stress alteration,
dynamic unclamping on extensional sides and clamping on
compressional sides of bulging bumps on the fault surface.
Nonetheless, different friction laws provide different description
of the yielding process in response to stress heterogeneity
induced by non-planar geometry, resulting in different level of
geometrical impact on rupture propagation. Compared to the
model with SW, those models with rate- and state- friction at low
velocity levels (RS-A and RS-S) show geometrical effects to a
lesser extent, while the one with RS-FH inducing pulse-like
rupture instead of crack-like rupture shows much stronger
geometrical impacts on rupture propagation.
How stressed are we really? Harnessing community
models to characterize the crustal stress field in Southern
California, Karen M Luttrell (Oral Presentation 9/12/16 14:30)
The in situ crustal stress field fundamentally governs, and is
affected by, the active tectonic processes of plate boundary
regions, yet questions remain about the characteristics of this
field and the implications for active faults in the upper crust. We
investigate the nature of this stress field in southern California by
combining observations from the SCEC Community Stress
Model, including stress orientation, stress from topography, and
stress accumulation rate on major locked faults. First, we
estimate the magnitude of the non-lithostatic in situ stress field in
southern California by balancing in situ orientation indicated by
earthquake focal mechanisms against the stress imposed by
topography, which tends to resist the motion of strike-slip faults.
Our results indicate that most regions require in situ differential
stress of at least 20 MPa at seismogenic depth. In the areas of
most rugged topography along the San Andreas Fault System,
differential stress at seismogenic depth must exceed 62 MPa
consistent with differential stress estimates from complimentary
methods. Second, we assess the origin of the heterogeneity
observed in the stress field by combining stress accumulation on
major locked fault segments with stress from topography and a
simple 2-D tectonic driving stress. Our results suggest that in situ
stress heterogeneity at the regional scale is more influenced by
deep driving processes acting on a laterally heterogeneous crust
than by perturbations to the stress field associated with major
locked faults in the upper crust. Finally, we discuss some
potential avenues for moving toward a 4D representation of the
crustal stress field in southern California.
Brittle to ductile transition in a model of sheared granular
materials, Xiao Ma, and Ahmed E Elbanna (Poster Presentation
042)
Understanding the fundamental mechanisms of deformation and
failure in sheared fault gouge is critical for the development of
physics-based earthquake rupture simulations that are becoming
an essential ingredient in next generation hazard and risk
models. To that end, we use the shear transformation zone (STZ)
theory, a non-equilibrium statistical thermodynamics framework
to describe viscoplastic deformation and localization in gouge
materials as a first step towards developing multiscale models for
earthquake source processes that are informed by high-
resolution fault zone physics.
The primary ingredient of the STZ theory is that inelastic
deformation occurs at rare and local non-interacting soft zones
known as the shear transformation zones. The larger the number
of these STZs the more disordered (the more loose) the layer is.
We will describe an implementation of this theory in a 2D/3D finite
element framework, accounting for finite deformation, under both
axial and shear loading and for dry and saturated conditions. We
examine conditions under which a localized shear band may form
and show that the initial value of disorder (or the initial porosity)
plays an important role. In particular, our simulations suggest that
if the material is more compact initially, the behavior is more
brittle and the plastic deformation localizes with generating large
strength drop. On the other hand, an initially loose material will
show a more ductile response and the plastic deformations will
be distributed more broadly. We will further show that
incorporation of pore fluids alters the localization pattern and
changes the stress slip response due to coupling between gouge
volume changes (compaction and dilation) and pore pressure
build up. We validate the model predictions by comparing them
to available experimental observations on strain localization and
fault gouge strength evolution. Finally, we discuss the
implications of our model for gouge friction and dynamic
weakening.
Characterizing the Geometry and Seismotectonics of the
Hilton Creek Fault System, Kyle P Macy, Amber K Lacy,
Jason De Cristofaro, and Jascha Polet (Poster Presentation 223)
The Hilton Creek Fault System, in the Eastern Sierra Nevada, is
characterized as a single NW-striking, NE-dipping, normal fault
from Davis Lake to the southern rim of the Long Valley Caldera.
Within the Caldera, the fault splays into multiple, parallel,
subsidiary faults that spread across the Caldera floor toward the
resurgent dome. Historically low seismicity in the area increased
in May of 1980, after the occurrence of 4 large earthquakes. In
the following decades, a series of earthquake swarms occurred
throughout the area, many of which were on or near the Hilton
Creek Fault System. Small vertical offsets have been associated
with these events
We are conducting an interdisciplinary geophysical study of the
Hilton Creek Fault System to delineate the fault splays within the
caldera and compare and contrast the seismic behavior of those
fault splays in the north, with that of the southern part of the Hilton
Creek Fault System. Our investigation will include the generation
of maps of the variation in b-values along different portions of the
fault, ground-based magnetic field measurements, high-
resolution total station elevation profiles, structure-from-motion
derived topography, and an analysis of earthquake focal
mechanisms. Topographic profiles of approximately 1 km in
length, produced from total station measurements, show at least
three distinct fault splays within the caldera with vertical offsets
between 0.5 to 1 m. More detailed topographic mapping through
structure-from-motion techniques is planned in the near future
WELCOME
2016 SCEC Annual Meeting page 207
and is expected to highlight the full length of these splays. We
will show maps of b-values, magnetic anomalies, topography,
various models of the Hilton Creek Fault System and cross-
sections through focal mechanism and earthquake catalogs, with
the main goal of integrating these observations into a single fault
geometry model.
Discriminating Between Induced vs Tectonic Seismicity in
Intraplate Regions: the Contribution of the Long-Term
History of Fault Behavior, Maria Beatrice Magnani, Michael L
Blanpied, Heather DeShon, and Matthew Hornbach (Poster
Presentation 024)
Since 2009 there has been an increase in rate of seismicity in the
Central U.S. (CUS), a major fraction of which has been
associated with shale gas production and related wastewater
injection. The surge of seismicity has resulted in an increased
seismic hazard in the region, which relies on the discrimination
between seismic activity that is anthropogenically induced and
that arising from natural tectonic deformation. This
discrimination, however, is particularly challenging because
tectonic strain rates and natural seismicity rates are low in this
intraplate region, such that tectonically active faults may display
periods of quiescence that are long (100’s to 1000’s of years)
relative to the short (10’s of years) instrumental record. In
addition, causative faults are unknown with a poor surface
expression, both types of seismicity occur on or reactivate
ancient faults in the Precambrian basement, and the instrumental
seismic record is sparse. While seismicity provides information
about the most recent deformation short-term deformation
history of the involved faults, the long-term history is missing.
Seismic reflection data offer a means to interrogate the long-term
history of these faults, a history which can be discriminatory. In
this poster we present examples from two regions of the CUS.
The first region shows examples of tectonically active faults
within the northern Mississippi Embayment imaged by high-
resolution seismic reflection data. The faults deform Quaternary
alluvium, as well as underlying sediments dating from Tertiary
through Paleozoic, with increasing amount of deformation with
formation age, suggesting a long history of activity. The second
region shows examples from the North Texas basin, a region of
ongoing shale gas exploitation. Here, industry seismic reflection
data image basement faults showing deformation of the
Precambrian and Paleozoic sequences, and little to no
deformation of younger sequences. Specifically, any vertical
offsets in the post-Pennsylvanian formations are below the
resolution of the seismic data at these depths (~10 m), far less
than expected had these faults accumulated deformation over
the long term, as observed in the first region. Assuming that
current seismicity recorded in the North Texas Basin is
representative of past seismic sequences, a vertical offset equal
to or less than 10 m along the currently active faults implies a
minimum average recurrence interval of ~150,000 yrs, based on
average displacement/sequence derived from cumulative
seismic moment calculations. These exceptionally long intervals
are at odds with the increasing number of faults reactivated in the
North Texas basin since 2008, and suggest that the recent
seismicity in the North Texas basin is highly anomalous, and
therefore more likely induced than natural.
Paleoseismic investigation of the Rose Canyon fault
zone, San Diego, CA, Eui-jo D Marquez, Jillian M Maloney,
Thomas K Rockwell, Neal W Driscoll, Scott Rugg, and Jeff M
Babcock (Poster Presentation 086)
The Rose Canyon fault zone (RCFZ) bisects the City of San
Diego, the 8th largest city in the U.S., and represents a major
seismic hazard to the greater metropolitan area that includes
Tijuana and surrounding cities. Onshore studies have shown that
the RCFZ is a predominantly right-lateral, segmented, strike-slip
fault zone that is capable of producing a M6.9+ event. However,
little information exists on the timing of mid- to late-Holocene
earthquakes, although the fault is known to have ruptured in the
past few hundred years. Furthermore, rupture patterns between
fault segments are not well constrained. To improve seismic
hazard assessments of the San Diego region, it is vital to
reconstruct the long-term paleoseismic history of the fault zone
and to understand the rupture patterns between fault segments.
The Spanish Bight fault segment, which runs through the San
Diego International Airport and splays into San Diego Bay, is a
critical part of the RCFZ. San Diego Bay is a pull-apart basin
created by a step in the RCFZ where subsidence and Holocene
sedimentation record the paleoseismic history of the fault zone.
Additionally, the Spanish Bight fault potentially records changes
in fault character and slip history as the RCFZ approaches this
transtensional step-over. We interpreted high-resolution Chirp
sub-bottom data that was collected from within San Diego Bay
offshore from the San Diego International Airport. We also
extracted data from existing geotechnical reports, including cone
penetrometer tests, core logs, and radiocarbon dates collected
from beneath the airport. Together, these data will be used to
generate a detailed active fault map and to resolve a stratigraphic
framework between the airport and bay sediments. We will
compare these data with the paleoseismic records from other
onshore fault segments to identify rupture patterns and evidence
for past earthquake clustering or triggering.
Is the CFM5.0 an Improvement? Evidence from
Mechanical Models of the Western Transverse Ranges
Region, Scott T Marshall, Gareth J Funning, and Susan E Owen
(Poster Presentation 151)
The SCEC Community Fault Model (CFM) is a widely-used
product that has existed for well over a decade and has seen
numerous revisions. As with all community-derived products,
testing the accuracy and performance of the model is key to
further development. To evaluate the ability for the CFM
geometry to reproduce long term fault slip rates, we present
results from geodetically-constrained mechanical models of the
western Transverse Ranges, CA. To determine if the current
CFM5.0 represents an improvement over the previous CFM4.0
and to identify the most likely geometry for the Ventura-Pitas
Point fault system, we have created three models to compare to
geologic slip rate estimates. The first model is based on CFM4.0,
the second based on CFM5.0 with a flat ramp geometry for the
Ventura-Pitas Point fault, and third, a CFM5.0 model with the
Ventura-Pitas Point fault having a constant dip angle. To
evaluate the performance of these three models, we compare the
slip rates in the Uniform California Earthquake Rupture Forecast
version 3 (UCERF3) to model-calculated average slip rates of
each fault segment. We find that while all three models match the
UCERF3 data well, the CFM4.0 model tends to produce slip rates
that are skewed towards the fast ends of the geologic slip rate
WELCOME
2016 SCEC Annual Meeting page 208
ranges. Despite significantly different subsurface geometries, the
two CFM5.0 models both produce similar average reverse slip
rates for the Ventura-Pitas Point fault of 3.1 and 3.4 mm/yr for
the ramp and constant dip models, respectively. Analysis of the
model-predicted surface slip distributions show that the existing
geologic slip rate estimate of 4.4-6.9 mm/yr was likely made in a
zone of faster than average slip along the fault. Because slip
rates are spatially variable, we also compare each geologic slip
rate point estimate to the model computed value at the fault
element nearest to the study site. We find that when compared
to model results, many existing geologic slip rate study sites are
not likely to record average slip rates.
Geophysical characterization of twelve CSMIP station
sites in Riverside County, California, Antony Martin, Lauren
Demine, William Dalrymple, Nolan Leue, and David Carpenter
(Poster Presentation 252)
The California Geological Survey recently funded geophysical
characterization of 12 California Strong Motion Instrument
Program (CSMIP) stations in western Riverside County,
California. The purpose of the investigation was to develop a
shear wave velocity (Vs) profile to a depth of 40 m and an
estimate the time-averaged shear-wave velocity of the upper 30
meters (Vs30) at each site. The CSMIP stations are located in
rural, suburban, and urban environments. Geologic conditions
ranged from shallow to relatively deep sedimentary basins.
Surface wave techniques, including the active-source multi-
channel analysis of surface waves (MASW), and passive-source
array microtremor and horizontal-to-vertical spectral ratio
(HVSR) techniques, were used for site characterization. Because
many of the sites had significant noise from traffic on adjacent
roads, the MASW technique was only expected to be able to
image Vs to depths of 20 to 30 m. Array microtremor
measurements, primarily made using nested triangle or L-
shaped arrays, ensured that depth of investigation exceeded 40
m. The vertical component of Rayleigh waves were recorded
during all active and passive surface wave acquisition. However,
complex Rayleigh wave propagation at two sites necessitated the
acquisition of the horizontal (radial) component of the Rayleigh
wave and Love wave MASW data. This demonstrates the
importance of a flexible field approach, where multiple surface
wave methods are available to be deployed as necessary.
Depth of investigation of the combined active and passive
surface wave soundings was in the 40 to 100 m range. The
fundamental mode assumption was generally valid for modeling
Rayleigh and Love dispersion data acquired during this
investigation. However, at 3 sites it was necessary to model
Rayleigh wave data using an effective mode assumption to
simulate a jump from the fundamental to first higher mode at low
frequencies. Crystalline rock, which ranged in depth from about
10 to 65 m, was encountered within the depth of investigation at
7 of the 12 sites. The predicted HVSR peak from the Vs models,
based on the diffuse field assumption, was relatively close to the
observed HVSR peak at most sites where the depth of
investigation was sufficient to estimate bedrock depth. Vs30 at
the 12 sites ranged from about 270 to 550 m/s; half of the sites
classified as NEHRP Site Class D and the other half as Site Class
C.
Measuring aseismic slip through characteristically
repeating earthquakes at the Mendocino Triple Junction,
Northern California, Kathryn Materna, Taka'aki Taira, and
Roland Bürgmann (Poster Presentation 238)
The Mendocino Triple Junction (MTJ), at the transition point
between the San Andreas fault system, the Mendocino
Transform Fault, and the Cascadia Subduction Zone, undergoes
rapid tectonic deformation and produces more large (M>6.0)
earthquakes than any region in California. Most of the active
faults of the triple junction are located offshore, making it difficult
to characterize both seismic slip and aseismic creep. In this work,
we study aseismic creep rates near the MTJ using
characteristically repeating earthquakes (CREs) as indicators of
creep rate. CREs are generally interpreted as repeated failures
of the same seismic patch within an otherwise creeping fault
zone; as a consequence, the magnitude and recurrence time of
the CREs can be used to determine a fault’s creep rate through
empirically calibrated scaling relations.
Using seismic data from 2010-2016, we identify CREs as
recorded by an array of eight 100-Hz PBO borehole
seismometers deployed in the Cape Mendocino area. For each
event pair with epicenters less than 30 km apart, we compute the
cross-spectral coherence of 20 seconds of data starting one
second before the P-wave arrival. The most similar events (with
median coherence above 0.95 at two or more stations) are
selected as CREs and then grouped into CRE families. Each
family is used to infer a local creep rate. On the Mendocino
Transform Fault, we find relatively high creep rates of >5 cm/year
that increase closer to the Gorda Ridge. Closer to shore and to
the MTJ itself, we find many families of repeaters on and off the
transform fault with highly variable creep rates, indicative of the
complex deformation that takes place there.
Geologic framework of the El Casco 7.5’ quadrangle:
southwestern gateway to the San Gorgonio Pass knot in
the San Andreas Fault zone, Jonathan C Matti (Poster
Presentation 123)
The El Casco 7.5' quadrangle is located at the SW portal of San
Gorgonio Pass (SGP), where Quaternary contractional
deformation in SGP gives way to transtensional deformation
associated with the right-lateral San Jacinto Fault. Important
stratigraphic and structural elements include (1) sedimentary
units that formed during the last 7 million years or so and record
tectonic events within the San Andreas Fault (SAF) system, (2)
the westernmost extent of the contractional SGP Fault zone, (3)
multiple strands of the San Jacinto Fault zone, and (4) the San
Timoteo Anticline, a NW-plunging, SW-vergent fold that deforms
the Neogene succession.
The upper Miocene to middle Pleistocene sedimentary sequence
is divided into three major units: (1) The Mt. Eden formation, (2)
the overlying San Timoteo formation, and (3) sedimentary
deposits of Live Oak Canyon (formerly interpreted as the upper
member of the San Timoteo formation). The sedimentary
sequence has the following significant attributes:
(1) Upper Miocene (~7 to 5 Ma) sediment in the Mt. Eden
formation was sourced from local Peninsular Ranges-type rocks,
and was dispersed radially into adjacent alluvial environments.
WELCOME
2016 SCEC Annual Meeting page 209
(2) Plio-Pleistocene (~5 to ~1.3 Ma) sediment in the San Timoteo
formation was sourced from San Gabriel Mountains-type rocks
NW of the El Casco quad and dispersed SE on a braidplain that
lapped onto Peninsular Ranges landscapes.
(3) At ~1.3 Ma the evolving San Timoteo Anticline disrupted the
San Timoteo depositional regime and sedimentary deposits of
Live Oak Canyon began to accumulate in a synformal basin NE
of the anticline and its ridge-like landscape; sediments of the Live
Oak Canyon sequence buttressed unconformably against the
anticlinal landform. Live Oak Canyon beds contain vertebrate
fossils of the Shutt Ranch local fauna (~900 ka) and span the
Brunhes-Matuyama magnetic reversal (780 ka).
These three Neogene depositional regimes correspond to
sequential tectonic phases of the SAF system: (1) the Mt. Eden
cycle (pre-5 Ma) corresponds with the Banning/San Gabriel
Fault; (2) the San Timoteo cycle (5 Ma to ~1.3 Ma) corresponds
with the San Andreas Fault; (3) the Live Oak Canyon cycle
corresponds with a tectonic event that disrupted the San Timoteo
braidplain, initiated the San Timoteo Anticline, and led to the
synformal depositional basin NE of the anticline. This outcome
most likely can be attributed to initiation of the San Jacinto Fault
as dextral slip stepped left from the SAF. Morton and Matti (1993)
suggest a 1.5 Ma age for this event; Matti and Morton (1993)
favor a 1.2 Ma age. Either scenario suggests that the San Jacinto
Fault initiated no earlier than 1.5 to 1.2 Ma, which yields a long-
term average slip rate between 16.7 mm/yr and 20.8 mm/yr
(based on total fault displacement of ~25 km). Kendrick and
others (2002) proposed a similar long-term average slip rate (>20
mm/yr) based on dated surfaces in the San Timoteo Badlands.
Development of extensional step overs within anisotropic
systems: Implications for the Rodgers Creek-Hayward
step over, Jessica A McBeck, and Michele L Cooke (Poster
Presentation 059)
The Rodgers Creek-Hayward extensional step over, within the
San Pablo bay in northern CA, USA, is estimated to pose one of
the highest likelihoods of rupture in northern CA. To better
constrain the fault geometry of the step over at seismogenic
depths, we simulate the propagation, interaction and linkage of
the Hayward and Rodgers Creek faults with the modeling tool
Growth by Optimization of Work (GROW). GROW simulates fault
development by adding radial elements to growing fault tips that
optimize the change in external work on the system. New GROW
implementations allow the simulation of fault development within
heterogeneous and anisotropic host rock, which facilitates
simulating Rodgers Creek-Hayward fault development within the
region’s variably deformed and metamorphosed detrital
sedimentary rocks dominated by serpentinite. Here, we construct
propagation forecast envelopes to reveal the extent of highly
efficient propagation paths that deviate from the optimal due to
heterogeneous weaknesses. The predicted forecast envelopes
and optimal paths produce step over geometries similar to
geometries observed in physical and numerical experiments, and
inferred from field observations. We also simulate the strength
anisotropy with an internal friction of the host rock that varies
non-linearly with orientation. Four parameters define this
anisotropy: the maximum internal friction of the host rock, the
orientation at which the internal friction is lowest, θmin, the value
of the minimum internal friction, μmin, and degrees from θmin at
which the internal friction reaches the maximum, θsat. To
investigate the sensitivity of fault propagation path to anisotropy
we vary θmin, μmin and θsat with GROW simulations of a simple
extensional step over. In these simple models, the faults are
planar and the separation between the faults approximates the
separation observed in geophysical imaging of the Rodgers
Creek-Hayward step over. Here, θmin and μmin exert a first-
order control on fault propagation path, and the tightness of the
weak orientations in the host rock, as parameterized with θsat,
exerts less influence. θmin orientations that deviate from the
existing faults and cross the extensional step over promote fault
linkage. Decreasing μmin promotes fault linkage because less
frictional work is required to grow the fault along its most efficient
path. GROW simulations with initial fault geometries that more
closely approximates the observed geometry reveal that the
predicted Hayward fault propagation path more closely matches
the geophysical data in models with a shorter Rodgers Creek
fault, in which the initial fault segments are underlapping.
Steady and time-dependent strain rate maps of California
from inversion of GPS time series, Robert McCaffrey (Poster
Presentation 142)
The GPS displacement time series from the Plate Boundary
Observatory (PBO), Crustal Motion Model 4 (Shen et al., 2011),
the Pacific Northwest (McCaffrey et al., 2013), University of
Nevada Reno (UNR) and others are used to estimate time-
dependent deformation of California from 1992 to the present
(2016.5). The time series are initially fit by estimating the
parameters of transient sources and a steady slope (site
velocity). The sources include all large earthquakes since 1992
plus after-slip for many of them and several volcanic sources.
The parameters are estimated by least-squares fit to the time
series. (Also estimated are seasonal terms and offsets due to
equipment changes and other non-tectonic causes.) Compared
to other methods where earthquake offsets are estimated
independently at each site and component, this approach
requires the offsets to be derived from just a few source
parameters and hence imposes a spatial correlation on the
offsets (and therefore on the site velocities). This helps improve
the estimates of velocities at sites with few observations, like
campaign or short-duration continuous sites. Combined with a
smoothed grid model to represent the steady motion (based on
the estimated time series slopes), both the steady and time-
dependent deformation can be estimated. This deformation
model is used to calculate strain rates and Coulomb stress
changes anywhere and at any time in the model domain. This
work is funded by SCEC and supports its Community Geodetic
Model effort.
Work in progress to estimate a Latest Pleistocene slip rate
for the Banning Strand of the San Andreas Fault near
North Palm Springs, Sally F McGill, Paula M Figueiredo, and
Lewis A Owen (Poster Presentation 125)
Existing late Pleistocene slip rate estimates for the southern San
Andreas fault reveal high rates of slip on the Mission-Mill Creek
strand in the Indio Hills but very low rates on that strand farther
northwest in the San Bernardino Mountains. The northwestward
reduction in slip rate on the Mission-Mill Creek strand is likely
accommodated by transfer of slip both northward to the Eastern
California shear zone and southwestward to the Banning and
WELCOME
2016 SCEC Annual Meeting page 210
Garnet Hill strands of the San Andreas fault. To better
understand the likely rupture paths and slip distribution in future
earthquakes on the southern San Andreas fault, it is important to
know how the latest Quaternary slip rate has been partitioned
among these fault strands. The published late Holocene slip rate
of the Banning strand is relatively low, suggesting that it has
played a relatively minor role in plate boundary slip over that time
period. We present here our work in progress at a new slip rate
site on the Banning fault near North Palm Springs, which will
provide an opportunity to confirm or modify that conclusion and
to test for variation in the slip rate of the Banning strand over time.
An ancient channel wall of Mission Creek that incised about 1.2
m depth into an older, distal alluvial surface has been offset 148
± 9 m by the Banning fault. In January 2016 we excavated three
dating pits at the site after a lengthy permitting process. Two 2-
m deep pits were excavated on the old alluvial surface into which
Mission Creek incised, and one 1.5-m deep pit was excavated
into the alluvial deposits that post-date the incision but that have
been offset out of the path of the modern channel of Mission
Creek. Twelve samples for optically stimulated luminescence
dating were collected (4 samples from each pit), as well as seven
samples for Be-10 depth profiles from each of the two pits on the
incised alluvial surface. Samples are being processed at the
Geochronology Laboratories at the University of Cincinnati and
preliminary results for some samples may be available in time for
the meeting. Six detrital charcoal samples from eolian and/or
fluvial overbank deposits in the top 0.5 m in the pits are all less
than 1000 years old. One detrital charcoal sample from 0.85 m
depth in a pit on the incised surface south of the fault has a late
Holocene age (2430 ± 25 radiocarbon years BP). The young age
of this sample suggests that it must also be from fluvial overbank
deposits that post-date the incision of Mission Creek into the
underlying, older alluvial surface. Luminescence dates from well
consolidated alluvial deposits near the base of the two deepest
pits are expected to provide a maximum age for the incision of
the offset channel wall and a minimum slip rate for this section of
the Banning fault.
Observations of the Temporal Evolution of Earthquake
Ruptures, Men-Andrin Meier (Poster Presentation 210)
Source time functions (STFs) of individual earthquakes show an
enormous variability. Yet, the large amount of earthquake source
data available today allows extracting common features among
groups of earthquakes, and to compare the "typical" behavior of
different groups. We use an extensive data set of local seismic
waveform data from shallow crustal earthquakes (a simple proxy
for STFs) as well as a compilation of STFs from teleseismic
source inversions from large subduction zone thrust earthquakes
(Ye et al., 2016) to compare the behavior of events with different
sizes (magnitude range M4.0-M9.1). We compare the observed
source behaviors against the predictions of conceptual
earthquake rupture models. We find that the STF growth rates
from both local and teleseismic data are significantly lower than
predicted by standard constant rupture velocity / constant stress
drop models (STF~t^2). Furthermore, the observations in both
data sets preclude a strongly deterministic rupture behavior: The
STFs of smaller and larger earthquakes are statistically
indistinguishable over roughly a third of their respective rupture
durations, implying that event initiation does not dictate final
rupture size. At the same time the statistically significant
differences for the following two thirds of rupture imply that the
preceding rupture evolution does condition the future rupture
evolution to some extent. While this "limited rupture predictability"
may be of limited practical use (in the sense of EEW-systems) its
observation places important constraints on the factors that
determine rupture propagation and termination.
Slip variability and temporal clustering along the Imperial
fault at Mesquite Basin, Imperial Valley, California, and
possible through-going rupture to the San Andreas fault,
Aron J Meltzner, Thomas K Rockwell, Rebecca Y Tsang, and
Paula M Figueiredo (Poster Presentation 124)
Paleoseismic trenches across the Mesquite Basin section of the
Imperial fault revealed several channels that cross the fault at a
high angle and that are displaced in the subsurface. These
channels incised into and are embedded within lacustrine strata
associated with filling events of Lake Cahuilla. Three-
dimensional excavation of these channels has yielded
information on slip in the past six surface ruptures. Displacement
is well documented for the 1940 and 1979 events, with 15–20 cm
of coseismic lateral slip occurring in each event at the site. A
small rill, which also corresponds to the feeder channel for older,
buried beheaded channels, is deflected by ~60 cm, which we
attribute to the 1940 and 1979 events plus creep and afterslip.
In contrast to the modern deflected channel, two subsurface
channels, each measuring ~50 cm in width, are completely
beheaded by slip on the fault, with no evidence for rounding of
the channels or flow along the fault. This relationship argues that
displacement in each of the corresponding prehistoric events—
the sixth and fifth events back (E6 and E5)—exceeded 50 cm;
the channel spacing suggests displacements of 1.4–1.5 m in
each of the these events. A younger channel complex contains a
sequence of nested channels that suggest two surface
ruptures—the fourth and third events back (E4 and E3)—each
with smaller displacements than in E6 and E5.
The ages of these past surface ruptures are constrained by their
stratigraphic relationship to lacustrine intervals, by local 14C
dating, and by analysis of the regional late Holocene lacustrine
stratigraphy. The youngest pre-1940 event (E3) occurred when
the site (at –32 m elevation) was underwater, likely around the
time of the most recent Lake Cahuilla highstand, AD 1711–1714.
Events E4–E6 must all have occurred between then and ca. AD
1500. The oldest channel, beheaded by E6, is offset a total of
~5.0 m from the active feeder channel. Hence, the northern
Imperial fault has sustained ~5.0 m of slip in the past 500 years,
with the majority (~4.4 m) occurring in the four earlier events
between about AD 1500 and 1714.
These results imply that: (1) the 1940 and 1979 displacements
were not “characteristic” for the northern Imperial fault; (2) slip
per event at the site ranges from about 0.15 to 1.5 m; (3) the slip
rate for the Imperial fault in Mesquite Basin is about 1 cm/yr for
the past five centuries, but the rate is necessarily higher near the
International Border; and (4) some paleoseismic displacements
are larger than expected for a site so close to the northern
terminus of the fault, suggesting that those ruptures may have
extended beyond the Imperial fault. Specifically, the size of two
large paleoseismic displacements on the Imperial fault and the
similarity of their timing with the timing of the two most recent
WELCOME
2016 SCEC Annual Meeting page 211
events on the southern San Andreas fault at Coachella argue that
through-going rupture scenarios must be considered.
Dynamic Triggering and fault geometry in the Central
Himalaya Seismic Gap, Manuel M Mendoza, Abhijit Ghosh,
and S S Rai (Poster Presentation 248)
The devastating Mw 7.8 Gorkha earthquake in Nepal on 25 April
2015, reawakened attention to the high earthquake risk along the
Himalayan arc. As a result of the Gorkha event, it is important to
identify where stresses are building up along and near the Main
Frontal Thrust that could produce the next devastating
earthquake in this area. The “central Himalaya seismic gap” is
located in Uttarakhand, India, west of the Gorkha rupture area,
where a large (> Mw 7.0) earthquake has not occurred for over
the last 200 years [Rajendran, C.P., & Rajendran, K., 2005]. This
500 - 800 km long along-strike unruptured segment of Himalaya
is poorly studied mainly due to the lack of modern and dense
instrumentations. Here, we analyze a rich seismic dataset from a
dense seismic network consisting of 50 broadband stations, that
operated between 2005 and 2012. We develop a robust
earthquake catalog of ~ 4,000 local events containing precise
locations, and magnitudes. By refining those locations in
HypoDD to form a tighter cluster of events using relative
relocation, we illustrate the fault geometry in this region with high
resolution. Furthermore, we observe periods of dramatic
increases in local micro-seismicity dynamically triggered by
remote, large teleseismic events. Generating a complete and
consistent earthquake catalog not only sheds new light on the
physical processes controlling the earthquake cycle in this locked
segment of Himalaya, it also illustrates how stress is transferred
along and between faults, whether via a static or dynamic
mechanism. From this catalog, we aim to reveal fault structure,
study seismicity patterns, and assess the potential seismic
hazard in the central Himalaya seismic gap, and the surrounding
region.
Towards automated estimates of directivity and related
source properties of small to moderate earthquakes in
Southern California with second seismic moments, Haoran
Meng, Yehuda Ben-Zion, and Jeff J McGuire (Poster
Presentation 229)
We develop a method for automated estimation of directivity,
rupture area, duration, and centroid velocity of earthquakes in
Southern California with second seismic moments. A 1D ray
tracing and an automated picking algorithm are combined to give
initial S phase picks. These are refined for time domain
deconvolution using a grid search within a short time window
around the automated S picks. Source time functions of target
events are derived using deconvolution with stacked Empirical
Green’s Function (EGFs) selected by spatial and magnitude
criteria as well as performances in the deconvolution. The use of
stacked EGFs helps to reduce non-generic source effects such
as directivity in individual EGFs. The method is suitable for
analysis of large seismic dataset and so far works for target
events with magnitudes as small as 3.5. Applications to several
small to moderate earthquakes in southern California indicates
that most have significant directivity. In particular, the recent June
2016 Mw 5.2 Borrego Springs earthquake is found to have strong
directivity to the northwest.
Geometric, kinematic, and temporal patterns of
Quaternary surface rupture on the Eastern Pinto
Mountain fault zone near Twentynine Palms, southern
California, Christopher M Menges, Jonathan C Matti, and
Stephanie L Dudash (Poster Presentation 108)
Detailed geologic mapping along the eastern Pinto Mountain
fault zone (PMfz) in the Twentynine Palms area reveals
previously unrecognized time-space variations in internal
geometry, kinematics, and Quaternary surface-rupture timing;
these variations appear to reflect fault-zone position relative to
intersections with transversely-oriented faults at both ends of the
mapped zone. The PMfz is an approximately E-W-trending, 80-
km-long left lateral fault that forms the structural boundary
between domains of transverse NW-trending right-lateral faults
in the central Mojave Desert to the north and subparallel E-W-
trending left-lateral faults in the eastern Transverse Ranges to
the south. Existing data suggest that most mid- to late
Quaternary activity is concentrated on sections of the PMfz west
of the Twentynine Palms area. However, our mapping along a
10-km section near Twentynine Palms indicates that the overall
zone of faulting increases in complexity and width to the east in
conjunction with deflection of the average fault zone orientation
from an E- to a more SE-direction. Eastward across this 10-km
section, the zone of surface ruptures increases in width from <
150 m at the W-end to a >1.5 km near the intersection with the
Mesquite Lake transverse fault on the eastern edge. This width
increase occurs via an series of E-directed branches and (or)
right steps in the composite fault zone that produce a new set of
ESE-oriented strands which diverge from the more E- to NE-
continuations of the strands entering the bifurcations from the
west. Geometrically, the composite fault zone consists of multiple
internal zones of clustered faulting, <5 m- to >200 m in width and
increasing from one to 4-5 in number eastward across the map
area, which are separated by relatively undeformed intervening
blocks. Each fault-rich domain comprises a complex array of
subparallel to branching discrete multiple component zones of
shears and fractures, 2- 8+ in number and centimeters to meters
in width, that collectively suggest wide zones of distributed
surface rupture. The ages of most recent ruptures--constrained
by fault-zone stratigraphy and scarps—vary across components
of the composite fault zone. Our youngest age estimates include:
(1) late Pleistocene rupture ages on the two main northern zones;
(2) multiple (minimum 2-3) Holocene-age ruptures, commonly as
young as late Holocene, on the two southern zones. Many
internal fault strands in the PMfz display geomorphic and
structural features (vertical scarps, fault-bounded ridges and
popup structures, oblique slickenlines, net displacement
estimates, and folded or tilted strata) indicative of secondary
transpression, commonly localized at SE-trending right-shifted
constraining bends, branches, or step-overs on internal fault
zones; this deformation increases eastward in intensity and
coverage as the zones approach transverse-fault intersections.
Seismic and Aseismic Moment Budget and Implication for
the Seismic Potential of the Parkield Segment of the San
Andreas Fault, Sylvain G Michel, Jean-Philippe Avouac,
Romain Jolivet, and Lifeng Wang (Poster Presentation 305)
This study explores methods to assess the seismic potential of a
fault based on geodetic measurements, geological information of
WELCOME
2016 SCEC Annual Meeting page 212
fault slip rate and seismicity data. The methods are applied to the
Parkfield's section along the San Andreas Fault at the transition
zone between the SAF creeping segment in the North and the
locked section to the south, where a Mw~6 earthquake has
occurred every 24.5 years on average since the M7.7 Fort Tejon
event in 1857. We compare the moment released by all the
known earthquakes and associated postseismic deformation
with the moment deficit accumulated during the interseismic
period. We find that the recurrence of M6 earthquakes is
insufficient to close the slip budget and that larger events are
probably needed. We will discuss and evaluate various possible
scenarios which might account for the residual moment deficit
and implications of the possible magnitude and return period of
Mw>6 earthquakes on that fault segment.
Aerial2lidar3d: A New Point Cloud-Optical Image
Matching Technique to Quantify Near-Field, Surface Co-
Seismic Deformation in 3D: Application to the 1999 Mw
7.1 Hector Mine Earthquake and New Surface Offset
Measurements of the 1999 Mw 7.6 Izmit earthquake, Chris
W Milliner, James F Dolan, Robert Zinke, James Hollingsworth,
Sebastien Leprince, and Francois Ayoub (Poster Presentation
088)
Measurements of co-seismic surface deformation of large
magnitude earthquakes are of critical importance for the
characterization of distributed deformation and understanding
fault zone mechanics. Recent developments in geodetic
matching techniques such as lidar differencing have proven vital
in providing accurate measurements of the near-field, co-seismic
deformation patterns of surface ruptures, areas where InSAR
typically decorrelates and traditional field surveys lack
measurement of distributed deformation. However, a major
limitation to the differencing approach is that owing to the
expense and recent development of lidar it’s spatial coverage is
largely limited, and most legacy data is of poor quality. Here we
develop a new and complementary technique to quantify the full
3D pattern of co-seismic surface deformation by matching post-
event lidar data with pre-event DEMs generated from stereo-pair
air photos, data that is inexpensive and ubiquitous in both space
and time, thus resolving the issue of missing pre-earthquake lidar
data anticipated in measuring future events. We apply this new
technique to the 1999 Mw 7.1 Hector Mine earthquake, using
Structure for Motion software (to build the topographic point
cloud), ICP (to co-register and match the point clouds), and
COSI-Corr (to quantify the horizontal motion). Using this
approach we successfully produced detailed 3D deformation
maps at 15 m resolution, detecting 0.2-1.35 m of vertical fault slip
along a 3.5 km stretch of the Hector Mine surface rupture. In
addition, we also present new fault offset measurements along
the 1999 Mw 7.6 Izmit surface rupture using high resolution air
photos, measured from offset tree lines, fences and roads, and
comparing these to field survey measurements to assess
whether they have missed distributed, ‘off-fault’ deformation.
Detailed measurements of co-seismic deformation using high-
resolution optical images has implications for better
understanding fault segmentation and linkage, important
information for PSHA models, constraints of surface slip in finite
slip inversions, and reliable data to accurately constrain empirical
scaling relations.
SCEC-VDO Reborn: A New 3-Dimensional Visualization
and Movie Making Software for Earth Science Data, Kevin
R Milner, FNU Sanskriti, John Yu, Scott Callaghan, Philip J
Maechling, and Thomas H Jordan (Poster Presentation 335)
Researchers and undergraduate interns at the Southern
California Earthquake Center (SCEC) have created a new 3-
dimensional (3D) visualization software tool called SCEC Virtual
Display of Objects (SCEC-VDO). SCEC-VDO is written in Java
and uses the Visualization Toolkit (VTK) backend to render 3D
content. A prior version of SCEC-VDO, which began
development in 2005 under the SCEC Undergraduate Studies in
Earthquake Information Technology internship, used the now
unsupported Java3D library. Replacing Java3D with the widely
supported and actively developed VTK libraries not only ensures
that SCEC-VDO can continue to function for years to come, but
allows for the export of 3D scenes to web viewers and popular
software such as Paraview.
SCEC-VDO offers advantages over existing 3D visualization
software for viewing georeferenced data beneath the Earth’s
surface. Many popular visualization packages, such as Google
Earth, restrict the user to views of the Earth from above,
obstructing views of geological features such as faults and
earthquake hypocenters at depth. SCEC-VDO allows the user to
view data both above and below the Earth’s surface at any angle.
It includes tools for viewing global earthquakes from the U.S.
Geological Survey, faults from the SCEC Community Fault
Model, and results from the latest SCEC models of earthquake
hazards in California including UCERF3 and RSQSim. Its object-
oriented plugin architecture allows for the easy integration of new
regional and global datasets, regardless of the science domain.
SCEC-VDO also features rich animation capabilities, allowing
users to build a timeline with keyframes of camera position and
displayed data. The software is built with the concept of
statefulness, allowing for reproducibility and collaboration using
an xml file.
SCEC-VDO runs on all recent 64-bit Windows, Mac OS X, and
Linux systems with Java 8 or later. More information, including
downloads, tutorials, and example movies created fully within
SCEC-VDO is available here: http://scecvdo.usc.edu
Smartphone-Based Earthquake Early Warning in Chile,
Sarah E Minson, Benjamin A Brooks, Sergio Barrientos, Juan
Carlos Baez, Maren Boese, Todd L Ericksen, Christian
Guillemot, Elizabeth S Cochran, Jessica R Murray, John O
Langbein, Craig L Glennie, and Christopher Duncan (Poster
Presentation 192)
Chile faces a high seismic hazard with magnitude 7.5+
earthquakes occurring roughly every 1.5 years. Accordingly,
Chile has a need for both earthquake early warning and local
tsunami early warning. The Chilean Centro Sismológico Nacional
(CSN) is in the process of establishing an early warning system
with scientific grade instrumentation. In parallel, we are
developing and deploying an early warning system in Chile using
only smartphones and an inexpensive GNSS add-on to
incorporate SBAS corrections. The smartphone uses an Android
application to collect data from the smartphone’s onboard
accelerometer and the GPS chip. The application analyzes and
transmits relevant data to a central server where we use the
FinDer-BEFORES algorithm to detect earthquakes and produce
WELCOME
2016 SCEC Annual Meeting page 213
a real-time joint seismic-geodetic finite-fault distributed slip
model (for sufficiently large magnitude earthquakes) or a near-
field acceleration-based line source model (for smaller
magnitude earthquakes). Accurate ground shaking forecasts
could be provided by either earthquake source model, and
distributed slip models for larger offshore earthquakes could be
used to infer seafloor deformation and thus provide local tsunami
warning. Our goal is to build and deploy over 200 smartphone-
based monitoring stations this year. Although this project utilizes
the smartphone-based sensor as part of a fixed network, this
approach could also be implemented in a crowd-sourced
manner. In November, 2015, the first 8 sensor units were
installed in the region of the 2015 Mw 8.3 Illapel, Chile
earthquake. While the early warning system is not yet live,
retrospective batch processing of all the data collected from
these few sensors shows that our proposed analysis method
successfully detects, locates, and estimates the magnitude for
two Mw≥5.5 earthquakes that have occurred since the sensors
were deployed while producing zero false alarms. We expect the
system’s performance to improve significantly once the sensor
network is expanded beyond this nominal initial deployment.
Activity of the Mill Creek and Mission Creek strands of the
San Andreas fault through the San Gorgonio Pass region,
Alex E Morelan, Michael E Oskin, Judith S Chester, and Daniel
F Elizondo (Poster Presentation 131)
We present new observations that constrain the recent slip
history of the Mill Creek and Mission Creek strands of the San
Andreas fault. These faults are the northern strands of a complex
series of strike-slip and thrust faults through the San Gorgonio
Pass stepover, an important structural barrier that affects seismic
hazard in southern California. Understanding the recent activity
on each strand of this complex system will help define potential
paths for future large, throughgoing San Andreas fault ruptures.
In the Raywood Flat area, the Mill Creek fault cuts the base of
the upper Raywood Flat fill, a ~50 m thick package of debris-flow
deposits. The upper section of the Raywood Flat deposits,
however, are not cut by the fault. On the surface of this deposit,
a 15 m-wide channel, flanked by boulder-rich, debris-flow levees,
crosses the projection of the Mill Creek fault without evidence of
offset. We collected boulder-top samples for cosmogenic
exposure age-dating of these levees and present preliminary
results. Additionally, we mapped inset terraces along the incised
channel of the East Fork Whitewater River drainage that also do
not show evidence of fault offset, and we collected a depth profile
through the uppermost Raywood Flat fill in order to further assess
its age. To the south, across Yucaipa Ridge and between the Mill
Creek and Mission Creek strands, the basement rocks are cut by
an extensive system of smaller faults and associated distributed
shear zones that record a transfer of displacement from the Mill
Creek to Mission Creek fault systems. Along several transects
through this transfer system, east of Forest Falls, we also find no
evidence of recent movement. Along the Mission Creek strand,
newly devegetated B4 airborne lidar data reveal fault scarps
cutting across hillslopes and alluvial fans between the San
Bernardino strand and lower Raywood Flat for a distance of 4
km. We identified a lateral offset of 4-6 m in an alluvial fan deposit
within a tributary of Banning canyon, and sampled a suite of
boulders to estimate the age of this deposit. This site documents
that the Mission Creek fault is active and serves as a potential
rupture path through the San Gorgonio Pass, bypassing the
structural complexity of the San Gorgonio Pass thrust to the
south and the eastern portion of the Mill Creek fault to the north,
the latter of which appears to be inactive since the Latest
Pleistocene.
Ground motions from induced earthquakes in Oklahoma
and Kansas: implications for seismic hazard, Morgan P
Moschetti, Steven Rennolet, Eric M Thompson, William Yeck,
Dan McNamara, Robert Herrmann, Peter M Powers, and Susan
Hoover (Poster Presentation 271)
Recent efforts to characterize the seismic hazard resulting from
increased seismicity rates in Oklahoma and Kansas highlight the
need for a regionalized ground motion characterization. To
support these efforts, we measure and compile strong ground
motions and compare these average ground motion intensity
measures (IMs) with existing ground motion prediction equations
(GMPEs). IMs are computed for available broadband and strong-
motion records from M≥3 earthquakes occurring January 2009–
April 2016, using standard strong motion processing guidelines.
We verified our methods by comparing results from specific
earthquakes to other standard procedures such as the USGS
Shakemap system. The large number of records required an
automated processing scheme, which was complicated by the
extremely high rate of small-magnitude earthquakes 2014–2016.
Orientation-independent IMs (RotD50, RotD100) include peak
ground motions (acceleration and velocity) and pseudo-spectral
accelerations (5 percent damping, 0.1–10 s period). Metadata for
the records included relocated event hypocenters. The database
includes more than 160,000 records from about 3200
earthquakes. Estimates of the mean and standard deviation of
the IMs are computed by distance binning at intervals of 2 km.
Mean IMs exhibit a clear break in geometrical attenuation at
epicentral distances of about 50–70 km, which is consistent with
previous studies in the CEUS. Comparisons of these ground
motions with modern GMPEs provide some insight into the
relative IMs of induced earthquakes in Oklahoma and Kansas
relative to the western U.S. and the central and eastern U.S. The
site response for these stations is uncertain because very little is
known about shallow seismic velocity in the region, and we make
no attempt to correct observed IMs to a reference site conditions.
At close distances, the observed IMs are lower than the
predictions of the seed GMPEs of the NGA-East project (and
about consistent with NGA-West-2 ground motions). This ground
motion database may be used to inform future seismic hazard
forecast models and in the development of regionally appropriate
GMPEs.
Incorporation of Local Site Effects in Broadband
Simulations of Ground Motions: Case Study of the Wildlife
Liquefaction Array, Ramin Motamed, and John G Anderson
(Poster Presentation 281)
The principal objective of this research is to explore the most
efficient approach to incorporate local site effects in Broadband
(BB) simulations of ground motions and include the nonlinear soil
behavior in the overall response. In this research, we will utilize
ground motions recorded at the Wildlife Liquefaction Array (WLA)
in Southern California for the purpose of validation of this
research.
WELCOME
2016 SCEC Annual Meeting page 214
We will select at least 10 seismic events recorded by this array.
The selected data will be grouped into two categories: (1)
Category 1 which includes small-to-moderate ground shakings
which initiated negligible excess pore water pressure and limited
soil nonlinearity, and (2) Category 2 which consists of strong
motions produced significant excess pore water pressure and
substantial soil nonlinear behavior.
This study will explore different approaches in incorporating local
site effects in the BB simulations including:
1. Direct method: in this method, the subsurface soil properties
are explicitly incorporated in the BB simulations. An equivalent
linear approach is possible in this alternative, at least with the
Composite Source Model approach (Anderson, 2015) since this
model retains a full wave propagation solution at high
frequencies.
2. Two-stage simulation: consisting of the BB simulations for
predicting motions for some horizon beneath the surface and
then calculating the soil response using one of the Site Response
Analysis (SRA) programs such as Deepsoil (Hashash et al. 2015)
or D-MOD2000 (Matasovic and Ordonez 2011).
3. Transfer function method: because in many cases, less
information is available than what is known of WLA, our
experiences with the above approaches will be used to evaluate
the use of transfer functions, including phase, that can be applied
to surface synthetics.
We will evaluate the efficiency of different methods in inclusion
of local site effects in the BB simulations using recorded data at
WLA and quantify the quality of predictions using the Goodness-
of-Fit (GOF) concept proposed by Anderson (2004).
Comparison of Source Inversions and Stress Drops with
In-Situ Observations of Faulting, Pamela A Moyer, Margaret
S Boettcher, and William L Ellsworth (Poster Presentation 053)
A Mw 5.4 earthquake, which occurred on 5 August, 2014 beneath
the Moab Khotsong gold mine, in Ornkey, South Africa provides
a unique opportunity to directly compare seismological
inferences with in-situ observations of faulting. The Mw 5.4
earthquake sequence was recorded by a dense surface seismic
network, in-mine seismometers, and a strainmeter. In addition,
the Mw 5.4 earthquake is the primary target of the proposed
DSeis International Continental scientific Drilling Program
project. With abundant, high quality waveforms, in-situ stress
measurements, and fault core, this earthquake can be an
important test case for the Source Inversion Validation (SIV)
community.
Using the surface seismic network data, we have developed an
initial source model of the space and time evolution of rupture of
the Mw 5.4 Orkney earthquake. The inversion for a one-
dimensional velocity model for the rupture area will be obtained
using VELEST. The finite fault inversion will be formulized using
minimal assumptions and we will require the grid model to be
consistent with the frequency content of the data in both space
and time. The displacement seismograms will be modeled to
capture near field phases with a causality condition applied, but
no other constraints will be placed on the rupture evolution. We
will use our kinematic model to infer the frictional evolution during
rupture and an empirical green's function spectral ratio technique
to obtain corner frequency for stress drop of foreshocks and
aftershocks of the Orkney earthquake. Our results will be
compared with direct evidence of the fault zone material from
fault cores through the rupture surface and stress measurements
performed in and surrounding the fault zone.
AWP-ODC Open Source Project, Dawei Mu, Alex Breuer,
Josh Tobin, Hui Zhou, and Yifeng Cui (Poster Presentation 336)
AWP-ODC is widely used community software that simulates
large-scale 3D seismic wave propagation in a viscoelastic
medium. We have recently undertaken an open source release
of AWP-ODC-OS for the first time. This open source effort is
dedicated to developing high-quality software for regional
petascale earthquake simulations that is freely available to the
community. We have licensed the AWP-ODC-OS software under
BSD-2 and will provide continuous updates to users. Extensive
documentation has been prepared to describe how to install this
software, as well as how to prepare the source and mesh files as
input to execute this AWP-ODC-OS code.
AWP-ODC-OS is a parallel code implemented in C and CUDA
which obtains high portability, high performance and high
scalability. The GPU implementation provides frequency-
dependent visco-elasticity, free surface boundary conditions and
absorbing boundary conditions, as well as support for multi-node
simulations through the MPI protocol.
We will also present updated version of SEISM-IO, a modular I/O
library tailored to AWP-ODC-OS. This library simplifies the
programming of parallel I/O for code developers by hiding more
complicated aspects from the user. By adding a unified software
layer between application and file system, SEISM-IO is capable
of calling other high-performance, widely used I/O libraries like
HDF5, PNetCDF and ADIOS for lower level I/O operations.
SEISM-IO is a highly robust library, which has been tested with
realistic cases that require I/O with 24,000 CPU cores.
Holocene geologic slip rate for the Mission Creek fault at
the Three Palms site in the Indio Hills, Juan J Munoz,
Whitney M Behr, Warren D Sharp, Rosemarie Fryer, and Peter
O Gold (Poster Presentation 127)
Slip on the southern San Andreas fault in the northwestern
Coachella Valley in Southern California is partitioned between
three strands, the Mission Creek, Garnet Hill, and Banning
strands. In the vicinity of the Indio Hills, the NW striking Mission
Creek strand extends from the Indio Hills into the San Bernardino
Mountains, whereas the Banning and Garnet Hill strands strike
WNW and transfer slip into the San Gorgonio Pass region.
Together, these three faults accommodate ~20 mm/yr of right-
lateral motion. Determining which strand accommodates the
majority of fault slip and how slip rates on these strands have
varied during the Quaternary is critical to seismic hazard
assessment for the southern California region. Here we present
a new Holocene geologic slip rate from an alluvial fan offset along
the Mission Creek strand at the Three Palms site in the Indio
Hills. Field mapping and remote sensing using the B4 LiDAR data
indicates that the Three Palms fan is offset ~57 +/- 3 meters. U-
series dating on pedogenic carbonate rinds collected at 25-100
cm depth within the fan deposit constrain the minimum
depositional age to 3.49 +/- 0.92 ka, yielding a maximum slip rate
of 16 +6.1/-3.8 mm/yr. This Holocene maximum slip rate overlaps
WELCOME
2016 SCEC Annual Meeting page 215
within errors with a previously published late Pleistocene slip rate
of 12-22 mm/yr measured at Biskra Palms, a few kilometers to
the south. Cosmogenic 10Be surface exposure samples were
also collected from the fan surface to bracket the maximum
depositional age. These samples have been processed and are
currently awaiting AMS measurement.
Estimating secular velocities from GPS data
contaminated by postseismic motion at sites with limited
pre-earthquake data, Jessica R Murray, and Jerry Svarc
(Poster Presentation 154)
Constant secular velocities estimated from Global Positioning
System (GPS)-derived position time series are a central input for
modeling interseismic deformation in seismically active regions.
Both postseismic motion and temporally correlated noise
produce long-period signals that are difficult to separate from
secular motion and can bias velocity estimates. For GPS sites
installed post-earthquake it is especially challenging to uniquely
estimate velocities and postseismic signals and to determine
when the postseismic transient has decayed sufficiently to
enable use of subsequent data for estimating secular rates.
Within 60 km of the 2003 M6.5 San Simeon and 2004 M6
Parkfield earthquakes in California, 16 continuous GPS sites
(group 1) were established prior to mid-2001, and 52 stations
(group 2) were installed following the events. We use group 1
data to investigate how early in the post-earthquake time period
one may reliably begin using group 2 data to estimate velocities.
For each group 1 time series, we obtain eight velocity estimates
using observation time windows with successively later start
dates (2006 – 2013) and a parameterization that includes
constant velocity, annual, and semi-annual terms but no
postseismic decay. We compare these to velocities estimated
using only pre-San Simeon data to find when the pre- and post-
earthquake velocities match within uncertainties. To obtain
realistic velocity uncertainties, for each time series we optimize a
temporally correlated noise model consisting of white, flicker,
random walk, and, in some cases, band-pass filtered noise
contributions. Preliminary results suggest velocities can be
reliably estimated using data from 2011 to the present. Ongoing
work will assess velocity bias as a function of epicentral distance
and length of post-earthquake time series.
Assessment of Predictive Values of Site Response based
on GMPE approaches using a Large-N array, Nori Nakata
(Poster Presentation 277)
Ground-motion prediction is a key component for seismic hazard
analysis, which is typically carried out with ground-motion
prediction equations. The standard deviation of the best-fit model
characterizes the residual ground motion variability. These
residuals are an admixture of random variability (aleatory
uncertainty) and modeling error (epistemic uncertainty). When
we compute site-specific ground motion prediction, the residuals
can become smaller since different sources, paths, and sites are
not mixed. Recent development of dense, capable, and long-
term seismometer networks allows us to estimate the site-, path-
, and source-specific variability. In this study, we use a very
dense array (Large-N array) at Long Beach, California, and
demonstrate the potential to characterize the spatial variability for
the vertical component of ground motion. Because of the density
of the array (2500 receivers with 100-m spacing), we can
estimate very dense site information (i.e., detailed near-surface
velocities) from the ambient seismic field. With this velocity
model, we can theoretically estimate the site amplification. The
recording time is only a couple of months, but we observed more
than 10 earthquakes, and hence we can estimate site-, path-,
and source-specific variability of ground motion. Since we do not
have multiple events occurring in close proximity to one another,
we cannot separate the source-location and source-parameter
effects.
Hybrid Broadband Ground motion simulations of Porters
Pass fault earthquakes, Mohammad A Nazer, Hoby N
Razafindrakoto, and Brendon A Bradley (Poster Presentation
285)
We present ground motion simulations of the Porters Pass (PP)
fault in the Canterbury region of New Zealand; a major active
source near the Christchurch city. The active (slip rate 3.2-4.1
mm/yr) segment of the PP fault has an inferred length of 82 km
and a mostly strike-slip sense of movement. This slip makes up
approximately 10% of the total 37 mm/yr margin-parallel plate
motion whilst also comprising a significant proportion of the total
strain budget in regional tectonics. Given that the closest
segment of the PP fault is less than 45 km from Christchurch city,
the PP fault is crucial for accurate earthquake hazard
assessment for this major population center.
We employ the hybrid simulation methodology of Graves and
Pitarka (2010, 2015), which combines low (f<1 Hz) and high (f>1
Hz) frequencies into a broadband spectrum. We have used
validations from three moderate magnitude events (Mw4.6 Sept
04, 2010; Mw4.6 Nov 06, 2010; Mw4.9 Apr 29, 2011) to build
confidence for the Mw > 7 PP simulations. Thus far our
simulations include multiple rupture scenarios which test the
impacts of hypocenter location and the finite-fault stochastic
rupture representation of the source itself. In particular, we have
identified the need to use location-specific 1D VS/VP models for
the high frequency part of the simulations to better match
observations.
Effects of elastoplastic material properties on shallow fault
slip and surface displacement fields, Johanna M Nevitt,
Benjamin A Brooks, Todd L Ericksen, Craig L Glennie,
Christopher M Madugo, David A Lockner, Diane E Moore, Sarah
E Minson, and Kenneth W Hudnut (Poster Presentation 075)
Inadequate knowledge of fault slip and off-fault deformation
within the shallow crust (<1 km depth) limits our understanding
of fundamental fault physics and seismic hazard associated with
near-surface rupture. Recent technological advances in near-
field geodesy (e.g., mobile LiDAR), however, provide the
displacement data necessary to invert for shallow slip. Such
analysis will benefit from a deeper understanding of the
rheological properties of surface materials (e.g., soil, gravel) and
their effect on partitioning deformation between discrete fault slip
and distributed off-fault deformation. We investigate shallow fault
physics using co- and post-seismic observations from the Mw 6.0
2014 South Napa earthquake and mechanical models. Detailed
mapping reveals secondary echelon mixed-mode fractures
rotated ~20° clockwise from the West Napa fault, indicating the
primary fault did not rupture the surface. Still, the surface shear
zone recorded by displaced vine rows and imaged using mobile
WELCOME
2016 SCEC Annual Meeting page 216
LiDAR is remarkably narrow (<5 m), given that the site is
underlain by soil and a sedimentary basin up to 2 km deep. In
Abaqus, we build 3D finite element models that include a
sedimentary basin, defined by Drucker-Prager elastoplasticity,
above an elastic basement. Slip on a vertical fault through both
layers is governed by the Coulomb criterion. Material properties
(elastic moduli, cohesion, internal friction) are systematically
varied to assess their impact on the relationship between fault
slip and surface displacement. Preliminary results indicate that
increased cohesion, representing increased clay content,
promotes shear strain localization at the surface. Results from a
representative Napa model, based on lab testing of soil samples
and the best estimate of fault kinematics, are compared with
LiDAR data of displaced vine rows. The results have implications
for finite fault models, which may obscure near-surface slip by
neglecting plastic deformation in the shallow crust.
Evaluation of Site Response in the Los Angeles Basin
from Spectral Ratio Analysis of Microtremor Data from a
High Density Temporary Broadband Deployment,
Raymond Ng, and Jascha Polet (Poster Presentation 206)
Sedimentary basins, such as the Los Angeles basin, can
significantly amplify seismic ground motion and increase its
duration, considerably increasing the potential for damage and
loss in urban areas. Most site response studies in the Los
Angeles basin use strong ground motion data and thus require
local earthquakes. Another approach to investigate site response
is to apply techniques and methods developed for microtremor
recordings. We will present the results of our investigation of site
response within the Los Angeles Basin through the application of
the microtremor Horizontal-to-Vertical (H/V) spectral ratio
method using the Geopsy software. This method was applied to
3-component broadband waveforms from the Los Angeles
Syncline Seismic Interferometry Experiment (LASSIE). LASSIE
is a collaborative, dense array of 73 broadband seismometers
that were active for a two-month period from October until
November 2014, transecting the Los Angeles basin from Long
Beach to La Puente. The data from this array enabled us to make
measurements of small-scale lateral variations in the peak
frequency, amplitude, and directional dependency of the H/V
spectral ratio for a station separation of ~1 km across this highly
populated sedimentary basin. Data analysis and interpretation
were conducted in accordance with the Site Effects Assessment
Using Ambient Excitations (SESAME) guidelines. The frequency
and amplitude of peaks in microtremor spectral ratios have been
shown to be related to resonance frequency and site
amplification. Our results show long period spectral peaks at the
basin center at periods of 6 – 9.5 s and additional peaks in the
spectral ratio curves at much shorter periods for sites at the basin
edge. Long period spectral peak amplitudes range from 2 – 5.5,
with the highest values measured for the greater Long Beach
area. We observe directional dependency in the frequency and
amplitude of the long period spectral peaks in proximity to the
basin edge, which appears to correlate with the strike of the basin
structure. We will show profiles of the H/V amplitudes and peak
frequencies across the LA Basin and interpret our results in the
context of site response results from other studies, as well as
models of shallow and deeper basin structure, such as the SCEC
Unified Community Velocity Model (UVCM).
2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): Earthquake Magnitude
& Event Probability Forecasting, Daniel Ngu, Brady Guan,
Mingyan Zhou, Yipeng Li, Jozi K Pearson, Thomas H Jordan,
Kevin R Milner, and Mark L Benthien (Poster Presentation 331)
The 2016 Undergraduate Studies in Earthquake Information
Technology (USEIT) Probabilistic Forecasting Team created
forecasts of large events (magnitude 7+ earthquakes) that will
occur in the next week, month, year, and 10 years on the
southern San Andreas Fault (sSAF). We use the information from
the Uniform California Earthquake Rupture Forecast, Version 3
(UCERF3) report and synthetic catalogs of thousands of years of
seismicity from the Rate-State Earthquake Simulator (RSQSim)
to find the probability of such large events. Our first task was to
tackle a few questions using a previous RSQSim catalog which
simulated nearly one million years of seismic activity in California,
based on UCERF2 fault data. Using Open Seismic Hazard
Analysis (OpenSHA) tools with Java to analyze the catalog, we
created a script to process the data. These results would be used
to produce probabilistic statistics to aid with the creation of a
forecast. Afterwards, we modified the code to process the
catalogs from the High-Performance Computing (HPC) Team.
The resulting data was used to compare their many RSQSim
simulations to a UCERF3 report data set. We determined which
catalog had the least difference in the recurrence intervals to the
UCERF3 report. The selected catalog was then extended by the
HPC team and analyzed by us to estimate recurrence intervals,
participation probability, and magnitude probability density
distribution of earthquakes that occur on the sSAF, San Jacinto
Fault (SJF), and other related faults and region. Using Excel, we
graphed the data to visualize our data. Some other tasks we were
accountable for determining general probabilities in a set amount
of years, conditional probabilities of an earthquake on the sSAF,
conditional probabilities of two earthquakes happening on the
sSAF, and conditional and directional probabilities of an
earthquake occurring on the sSAF and SJF. The visualization of
our data is executed using Excel and SCEC-Virtual Display of
Objects (VDO), which displays earthquake occurrences through
an animation. Our results may be used by future SCEC interns
and employees, allowing them the option to either build on our
results or use it as a comparison datasheet.
Anomalous Uplift at Pitas Point: Implications from
Onshore & Offshore 3D Fault & Fold Geometry and
Observed Fault Slip, Craig Nicholson, Christopher C Sorlien,
and Tom E Hopps (Poster Presentation 006)
Based on four Recent ~8-m uplift events of coastal marine
terraces at Pitas Point, many believe these represent
earthquakes near M8 on the N-dipping Pitas Point-Ventura fault
(PPVF), part of the larger primarily offshore North Channel-Pitas
Point-Red Mountain fault system. However, this model of multiple
Holocene M8 events on the PPVF has major problems, not the
least of which are: failure of the 2D fold model used to properly
infer subsurface fault-fold geometry, an implied Holocene slip
rate for the blind PPVF that is inconsistent with offshore
observations, and the marked lack of near-surface fault rupture
or widespread tsunami deposits expected from such shallow
(<15 km depth) M8 events that would extend 10's of km offshore.
The reason for these discrepancies may be that uplift at Pitas
WELCOME
2016 SCEC Annual Meeting page 217
Point is driven primarily by slip on the S-dipping listric Padre Juan
fault (PJF), not solely by the PPVF. The PJF juxtaposes the
strongly N-verging San Miguelito anticline in its hanging wall
above the more symmetric Ventura Avenue-Rincon anticline in
its footwall. Fault and fold geometry is well determined by
industry wells that produce from the distinctly different upper San
Miguelito and lower Rincon oilfields, and by imaging offshore with
seismic reflection data. In the upper 3 km, the PJF exhibits up to
2.6 km of dip separation, in contrast to ~200 m of inferred
separation on the Ventura fault at similar depths. Much of this
PJF slip is syntectonic with growth of the Rincon anticline as PJF
splays are folded by this lower fold. The timing and slip involved
for San Miguelito fold growth and its emplacement against and
deformation of the lower Rincon anticline, the specifics of infered
Ventura fault propagation, and the geometry of the PJF and
PPVF requires that much of this fault slip occurred while the PJF
acted independently—not as a backthrust to the PPVF. A still
active listric PJF can help account for the observed uplift at Pitas
Point without recourse to M8 earthquakes. Regardless, its
presence helps explain why the uplift at Pitas Point is so
anomalous, and not necessarily indicative of the expected slip at
depth either along strike of the PPVF or the average slip during
large earthquakes. Rather, this uplift at Pitas Point is probably
localized to where slip on the PJF predominates, or where the
PJF and PPVF strongly interact, which limits the length & depth
of possible seismic ruptures and the geohazard & tsunami
potential of the active fault(s) involved.
Hybrid Genetic Deflated Newton Method for Distributed-
Source Optimization, Marcus M Noack, and Steven M Day
(Poster Presentation 349)
Earthquake fault parameter inversion is a promising field of
research. An accurately defined source leads in turn to accurate
ground motion predictions. The promising perspective comes at
a cost; local and global optimization procedures like the Newton
method and the genetic algorithm offer a trade-off between
accuracy and computational costs. Using the renowned Newton
method cannot lead to success because the misfit function to be
optimized exhibits many local optima because of the natural sine
shape if a wave. Furthermore, the global optimum might not be
unique. On the plus side, the Newton method is very efficient in
finding a stationary point.
As a global method, the genetic algorithm suffers from high
computational costs in high dimensional search spaces. Also,
since no local information of the function is used, the algorithm
cannot determine whether a stationary point is found. Therefore,
only one solution can be found. An advantage of the genetic
algorithm is, however, that the global optimum will be found
eventually. Genetic algorithm and Newton methods are only
examples of many local and global optimization methods but they
are representative in terms of their abilities.
To join the advantages of both methods, hybrid methods have
been developed. Hybrid methods are using a global optimization
scheme like the genetic algorithm to explore the parameter space
and a local scheme to find stationary points. One major drawback
persists; one optimum can be found several times which leads to
biased behavior and a premature convergence.
Recently, a new method with the name Hybrid genetic Deflated
Newton method (HGDN) has been proposed which uses the
advantages of a hybrid method. The new method places several
individuals randomly in the search space. All individuals perform
a Newton method to a stationary point. The found points are
transferred to a list and the points are removed from the function
by deflation. Afterwards, the entire population of individuals is set
back to its starting position and the search starts again. Already
found optima cannot be found again. Therefore, the individuals
converge in new stationary points. The procedure continues until
no more stationary points can be found in the vicinity. A genetic
algorithm replaces the fittest individuals and the procedure starts
over.
In this work we outline the work flow of using HGDN for fault
parameter optimization, illustrating the concept for the case of a
distributed acoustic source. Future work will consider the
corresponding elasto-dynamic source inverse problem.
2016 SCEC Undergraduate Studies in Earthquake
Information Technology (USEIT): SCEC-VDO
Development & Visualization, Rory C Norman, Kishan
Rajasekhar, Emy S Huang, Patrick Vanderwall, Allen Y Shi,
Rafael G Cervantes, Mark L Benthien, Thomas H Jordan, FNU
Sanskriti, Jozi K Pearson, John Yu, and Kevin R Milner (Poster
Presentation 332)
SCEC-VDO (Southern California Earthquake Center Virtual
Display of Objects) is a software program which allows users to
visualize different sets of earthquake data. SCEC-VDO is
developed by interns in Undergraduate Studies in Earthquake
Information Technology (USEIT), who add new functionalities to
it each year. This year, the SCEC-VDO team was tasked by
Thomas Jordan to revamp SCEC-VDO using The Visualization
Toolkit (VTK), which replaced the old unsupported Java 3D
platform, to visualize 30,000 year catalogues of multiple
earthquakes. To visualize these catalogues, a new open source
plugin was created to read data files provided by the High
Performance Computing (HPC) team, which then can display the
rupture events on SCEC-VDO. The most important improvement
made to SCEC-VDO is the complete makeover to the timeline
plugin and ability to save states of the program. The older version
of the timeline plugin was difficult to use. The new timeline allows
users to insert key frames which saves the state, such as the
camera position, the number of components on the map, the
transparency of the components, and the color of the
components. This makes it easy for users to make edits when
creating their video and renders stable videos, eliminating the
need for post video editing. The other plugins that were
developed include the Shake Map Plugin, Legend Plugin, and
the Hazus Plugin.
Beat the Quake: Designing an Earthquake Themed
Puzzle Room, Jenny Novak, and Julian C Lozos (Poster
Presentation 323)
California State University, Northridge (CSUN) is one of the
nation’s largest single-campus universities. The University is
located in a dense, urban setting with a history of exposure to
large earthquakes including the 1994 Northridge earthquake and
the 1971 Sylmar earthquake. With a majority of the current
student population born after 1994, the institutional memory of
historic resilience to these earthquakes is beginning to fade. In
order to better communicate earthquake risk and recommended
WELCOME
2016 SCEC Annual Meeting page 218
preparedness measures to current students, CSUN staff and
faculty collaborated to create a novel, interactive, educational
method. Beat the Quake is an earthquake themed puzzle room,
set in February 1994 just weeks after the magnitude 6.7
Northridge earthquake damaged the university. The puzzle room
challenges groups of 4-5 students to utilize teamwork to solve a
series of earthquake related puzzles before the earthquake hits,
simulated by the earthquake early warning sound. Puzzles
include identifying the San Andreas fault on a map, identifying
the magnitude of the 1994 Northridge Earthquake, utilizing
earthquake mitigation techniques to physically secure items in
the room, collecting preparedness items to build a kit, and
practicing “drop, cover, hold on” when the earthquake early
warning sounds. Beat the Quake debuted in April 2016 and
garnered 113 participants. A post-event survey revealed that
participants retained a high level of earthquake knowledge, with
over 98% correctly identifying the appropriate action to take
during an earthquake despite the fact that for most this was the
first disaster preparedness activity that they had participated in.
100% of those surveyed indicated that they would like to see the
Beat the Quake event held at CSUN on at least an annual basis.
Provably stable and high-order accurate finite difference
schemes on staggered grids and their potential
application for seismic hazard analysis, Ossian O'Reilly, and
Eric M Dunham (Poster Presentation 320)
When it comes to wave propagation over large distances, high-
order and centered finite difference approximations on staggered
grids are highly effective. For this reason, the 4th order staggered
grid scheme is used for many of SCEC’s high performance
computing efforts, ranging from large-scale earthquake hazard
simulations like ShakeOut to full-waveform tomography to
reciprocity-based CyberShake calculations. In these simulations,
it is essential to accurately model surface waves due to their
dominant role in ground motion at all but the closest distances.
However, unless care is taken in the implementation of the free
surface boundary condition, instability or inaccuracy can result. It
is common to decrease the order of accuracy near the free
surface or to introduce ad-hoc artificial dissipation or filtering to
avoid instability. Unfortunately, these techniques can severely
degrade the accuracy of the solution, reduce the maximum stable
time step, or involve additional parameters that can be
inconvenient for practitioners to tune. In this work, we construct
a provably stable numerical scheme on staggered grids that
retains high accuracy near boundaries. This is done by deriving
one-sided finite difference approximations near the boundary
that satisfy the principle of summation-by-parts (SBP). To
preserve the SBP property of the scheme, and establish stability
using energy estimates for the semi-discrete problem, boundary
conditions are weakly enforced. Here we demonstrate the
technique for the acoustic wave equation in Cartesian
coordinates, but its extension to the elastic wave equation is
anticipated to be straightforward. To handle complex geometry,
the technique can likely be extended to curvilinear grids, non-
conforming grids, or be coupled to other provably stable schemes
such as discontinuous Galerkin schemes.
A possible joint San Andreas-Imperial fault rupture in the
Salton Trough region, David D Oglesby, Christodoulos
Kyriakopoulos, Aron J Meltzner, and Thomas K Rockwell (Poster
Presentation 056)
We investigate the potential for large through-going ruptures
across the Salton Trough using both field data and numerical
simulations. Geological records from paleoseismic trenches
inform us of details of past ruptures (length, magnitude, timing),
while dynamic rupture models allow us to evaluate numerically
the mechanics of such earthquakes. The two most recent events
(Mw 6.9 1940 and Mw 6.4 1979) on the Imperial fault (IF) both
ruptured up to the northern end of the mapped fault, potentially
implying that ruptures must terminate there. This result is
supported by small displacements, ~20 cm, measured at the
Dogwood site near the end of the mapped rupture in each event.
However, 3D paleoseismic data from the same site
corresponding to the most recent pre-1940 event (1710 CE) and
the 5th (1635 CE) and 6th events back reveal up to 1.5 m of slip
in those events. Since we expect the surface displacement to
decrease toward the termination of a rupture, we postulate that
in these earlier cases the rupture propagated further north than
in 1940 or 1979. Furthermore, paleoseismic data from the
Coachella site (Philibosian et al., 2011) on the San Andreas fault
(SAF) indicates slip events ca. 1710 CE and 1588–1662 CE.
Thus, the timing of two large paleoseismic displacements on the
IF cannot be distinguished from the timing of the two most recent
events on the southern SAF. We investigate the possibility of
through-going rupture across the Salton Trough using 3D
dynamic finite element rupture modeling. In our work, we
consider two scenarios: rupture initiating on the IF propagating
northward, and rupture initiating on the SAF propagating
southward. Initial results show that in the former case, rupture
propagates north of the mapped northern terminus of the IF only
under certain pre-stress conditions, such as values of the seismic
strength parameter S = 0.45 to 2.0, and tends to stop for S = 2.5.
If rupture initiates on the SAF in the north, we find that it is easier
for it to propagate across the entire stepover region. The results
have implications for potential earthquakes in the region, with the
possibility of a preferred direction of rupture propagation through
the stepover.
Kumamoto earthquake: a complex earthquake sequence
with large strike-slip ruptures, Koji Okumura (Oral
Presentation 9/13/16 08:30)
The Mw 7.3 mainshock of the Kumamoto earthquake on April 16
ruptured the Futagawa fault zone mostly as it was forecasted.
However, the sequence of earthquakes, ruptures, and localized
severe shaking were beyond the forecast demonstrating the
complex nature.
The earthquake sequence began with a Mw 6.2 at 21:26 JST on
April 14. This one ruptured deeper part of the Takano and
Shirahata faults (TSF) in SW of the Futagawa fault. Strong
shaking (JMA scale 7 out of 7) damaged many houses in a small
area of Mashiki town-center to kill 9 people. Following another
Mw 6.0 on the April 14 source area, the Mw 7.3 occurred at 01:25
JST on April 16. The mainshock induced several Mw 5.3 to Mw
5.9 earthquakes outside its source area, at 75 km NW, 45 km
NW, and 40 km SW away from the epicenter. Those induced
earthquakes and aftershocks occurred in a zone about 120 km
long across Kyushu Island while the Mw 7.3 rupture was only 30
km long.
The 30 km long Mw 7.3 rupture mostly followed the previously
mapped Futagawa fault zone with dextral offset up to 2 m.
WELCOME
2016 SCEC Annual Meeting page 219
However, the surface offset of the TSF was only 0.3 m. Also, the
TSF ruptured twice both on April 14 and on April 16. Assuming
the two shocks ruptured the same fault plane, the small offset on
April 16 could be due to the strain release on April 14, or to the
rather short elapsed time of 1200 to 1500 years. The
northeasternmost 5 km section, which also caused severe
damages, cutting into the Aso caldera was not mapped before.
Rapid erosion in this high-relief area probably erased or buried
past surface ruptures. A 3.7 km long branch fault with ~1 m offset
appeared from a restraining bend of the Futagawa fault to the
Mashiki town-center with conjugate sinistral offset. These faults
ruptured active alluvial plain.
The severest damages were concentrated in a small area of the
Mashiki town-center where JMA intensity 7 was recorded twice
during Mw 6.2 and Mw 7.0. The possible causes of the severe
damages are the source process, site amplification, surface
faulting, slope failure and lateral spreading, and so on. It is
necessary to integrate analyses and models by different
disciplines. From geologic points of view, the presumable
complexity of shallow subsurface structures at the boundary
between Holocene graben fill in south and upland-forming
Pleistocene volcanics and gravels in north is a likely cause of the
localized amplification. Further investigation is necessary to
solve this significant hazard.
Rupture evolution and high-frequency radiation during
mega/large earthquakes, resolved by hybrid
backprojection technique, Ryo Okuwaki, and Yuji Yagi
(Poster Presentation 067)
Tracking spatio-temporal evolution of high-frequency (HF)
sources during rupture is a key to understand nature of
earthquake rupture process since HF radiation reflects an abrupt
change of rupture velocity and/or slip-rate. We have been
developing hybrid backprojection (HBP) technique; a scheme for
tracking HF sources at high resolution by stacking cross-
correlation functions of observed waveforms and theoretically
computed Green’s functions.
The HBP method overcomes a difficulty in ensuring a depth
resolution in projected images and mitigates signal-delayed
artifacts built in conventional backprojection (BP) methods due to
larger amplitude of reflected phases than that of direct P-phase
by using information of relative timing and amplitude between P-
phase and depth phases (pP, sP phases) in a cross-correlation
process. These features of the HBP method enable us to
integrate HF sources with slip distribution obtained by using the
waveform inversion and provide us more detailed view of rupture
evolution, e.g., cascading rupture or rupture-stopping
mechanism, which had been difficult to be considered with
conventional BP methods. In addition, an explicit use of Green’s
functions, which depend on depth and focal mechanism of
assumed source knot, can draw how geometric complexity in
fault system has an effect upon rupture propagation via tracking
HF sources with the HBP method.
In the meeting, we will present the recent observations of the
mega/large earthquakes (2008 Wenchuan China/2010 Maule
Chile/2015 Gorkha Nepal/2015 Illapel Chile) applying the HBP
method together with the waveform inversion, and discuss
diverse patterns of rupture evolution highlighting the
heterogeneous distribution of stress state or frictional property
along the fault, and the role of geometric barriers in rupture
propagation. Through applications of the HBP method, we have
encountered an artifact showing that the strength of HF signals
is systematically dependent on depth, and found that, in some
cases, rupture direction is not well resolved due to source-
receiver geometry. We will also present these problems we are
now facing, and explore some tentative ways to solve them.
Next Generation SDSU BBP Module Validation, Kim B
Olsen, and Rumi Takedatsu (Poster Presentation 001)
The SDSU module (BBtoolbox V1.5), merging low-frequency
deterministic signals with high-frequency scattering functions
(Olsen and Takedatsu, 2015), participated in and passed the
SCEC Broadband Platform validation exercise 1.0, where the
focus was on validating simulated median pseudo-spectral
accelerations (PSA) for large earthquakes in western and
eastern US and Japan, as well as NGA-west1 GMPEs. Here, we
have started to prepare the SDSU module for the next generation
seismic hazard analysis, by extending the validation to additional
metrics important for structural engineering, namely signal
duration, frequency-dependent ground motion variability (intra-
event standard deviation), and correlation of spectral
accelerations. Moreover, we ‘re-validate’ the SDSU module
against median NGA-west2 horizontal ground motion levels.
We find that the fit to durations depends strongly on the specifics
of how the values are calculated. For example, the SDSU module
provides generally good fit to durations for 9 western events if
evaluated by the current BBP alignment methodology for
goodness-of-fit calculation, starting at 5% energy level and
truncating the synthetic time series to fit the length of the data.
However, such truncation is not feasible without seismic records,
such as for seismic hazard analysis. Without the truncation, the
durations from the SDSU module are generally too long, and we
adjust the coda energy envelope of the synthetics to improve the
GOF for the durations. In general, intra-event standard deviations
for the Part A PSAs generated by the SDSU module using 50
realizations are in agreement with those from data. Preliminary
tests show that the PSA correlation structure for the SDSU
module deviates from data, generally too correlated for shorter
periods and with too small correlation for longer periods. We
propose a path toward better correlation structure for the SDSU
module for future work. Finally, we modify the module to fit
median PSAs according to NGA-west2 (SCEC Validation
Exercise, Part B). Among other differences compared to NGA-
West1, the NGA-West2 PSAs are generally lower for periods
between 0.2 s and 1 s. Through adjusted magnitude- and region-
dependent parameters for the shape of the source-time function
convolved with the high-frequency scattering functions we obtain
satisfactory fit for 50-realization median PSAs for Part B, mostly
with similar or smaller bias for the Part A scenarios.
Testing ETAS Catalogs from UCERF3, Morgan T Page,
Nicholas J van der Elst, Edward H Field, and Kevin R Milner
(Poster Presentation 299)
Epidemic-Type Aftershock Sequence (ETAS) catalogs
generated from the 3rd Uniform California Earthquake Rupture
Forecast (UCERF3) model are unique in that they represent a
marriage between two types of earthquake forecast models that
have never previously been formally combined. UCERF3 is a
complex fault-based model with geologic, paleoseismic, and
WELCOME
2016 SCEC Annual Meeting page 220
geodetic constraints as well as elastic rebound effects. The
ETAS model adds short-term spatio-temporal clustering from
aftershock triggering. Significant challenges in the construction
of the UCERF3-ETAS model were encountered, particularly with
respect to the magnitude-frequency distributions in the UCERF3
model and the application of elastic rebound.
Given the challenges present in its construction, to what extent
do UCERF3-ETAS catalogs reproduce the observed statistics in
real earthquake catalogs? We present Turing-style tests of the
temporal and spatial statistics of synthetic and real catalogs. We
find that UCERF3-ETAS is more spatially diffuse than the
observed historic catalog in California. This discrepancy is due to
the diffusivity of the background seismicity, which is based on a
model that has been quite successful in short-term tests – the
adaptive smoothed seismicity model of Helmstetter et al. (2007).
We also find that UCERF3-ETAS is lacking quiet periods that are
present in the real catalog. This is caused by the steady
background hum of the stationary events in the ETAS model,
which constitute 1/3 of all modeled earthquakes.
Characterization of fault motions observed with
UAVSAR., Jay W Parker, Andrea Donnellan, Scott Hensley,
and Margaret T Glasscoe (Poster Presentation 138)
Publically available data from the Uninhabited Aerial Vehicle
Synthetic Aperture Radar (UAVSAR), which measures
deformation between repeat visits, provide an excellent
opportunity for characterizing fault motions. These include the
main shock fault trace and near-field deformation, effects of
larger aftershocks and swarms, triggered slip at a distance, and
interseismic creep. Limitations include the following. The
discovery time to find events of interest in the large image data
set: the set today has nearly 2000 strips, each with roughly 100-
200 megapixels. Next, the common occurrence of incoherent
patches in an interferogram leaves irregular holes in the
deformation image. Next, the circumstance where faults with slip
much larger than a phase cycle or that cut across an entire
image: that defeats unwrapping. Finally, the overprinting of
deformation at spatial scales that match the signatures of
atmospheric water vapor irregularities and uncompensated
aircraft motion. To address these limitations, we demonstrate a
number of examples with remediation. The Geo-Gateway.org
visual interface to the UAVSAR data set allows fast browsing and
extraction of deformation profiles, allowing detailed examination
at full resolution before using the time and bandwidth required to
download images. We are also beginning to deploy automatic
surface fracture detection to highlight features of interest. We
have established reliable means of detecting and avoiding
incoherent patches and pixels so they do not interfere with
surface fracture detection and modeling. We have established by
far field considerations, the otherwise ambiguous large slip
across a rupture cutting entirely through an image using the
South Napa Earthquake, and established methods for estimating
a component of surface fault slip for each distinct view-direction
flight path. Two flight paths captured the La Habra Earthquake
coseismic deformation, allowing separate estimation of the
vertical deformation and one component of the horizontal
deformation. Ruptures that do not reach the surface but that
terminate within about 400 m, such as the Ocotillo aftershock of
the El Mayor-Cucapah Earthquake, are detected and given
rough characterization. Faults that appear to have very shallow
slip deficit, such as the South Napa event, display near-field
deformation that matches a simple deficit model.
Developing the Next Generation SCEC Web
Infrastructure, Edric Pauk, David Gill, Tran T Huynh, Mark L
Benthien, John E Marquis, Jason Ballmann, John Yu, and Philip
J Maechling (Poster Presentation 327)
The Southern California Earthquake Center's (SCEC) primary
mission is to gather data on earthquakes, integrate that data into
a physics-based understanding of earthquake phenomena, and
communicate that understanding to the society-at-large. One of
the key ways SCEC achieves its goals is through its online
presence. SCEC.org plays an important role as an information
hub for the earthquake science community and the scientific work
the community produces.
We have designed the new SCEC.org website to be a highly-
visible, curated archive of SCEC activities and accomplishments
that includes information about the wider earthquake science
community. Over the last year, our SCEC web development
group has developed and deployed a new SCEC.org website
based on the extensible, open-source Drupal Content
Management System (CMS). The widely-used Drupal software
system has a large and active developer community and thus
provides a modern, well-supported platform for SCECs future
web needs. Drupal helps to reduce the amount of custom
software we must develop to build and maintain our website.
Currently, SCEC.org web-based capabilities include the
presentation of recent SCEC results, meeting registrations,
proposal submissions, and abstract submissions. In the future,
we plan to extend SCEC.org web capabilities to include
distribution of multimedia content, mailing list management, and
discovery and access to SCEC research results. We believe the
new SCEC information systems infrastructure will help to
increase the quality and quantity of services we provide to the
community and will help the SCEC community accomplish our
mission.
Linking grain-scale and fault-scale earthquake
simulations, Ryan Payne, Benchun Duan, and David Sparks
(Poster Presentation 049)
Geologically active faults are complex structures, marked by an
interaction of many different physical processes at varying time
and length scales. Much progress has been made utilizing
advanced computer models as an effective way of studying the
different components of these systems, however, little has been
done to link various models together into a more complete and
realistic simulation of the natural world. This study utilizes a
dynamic rupture model along with granular dynamics simulations
to examine interactions between the two scales, particularly
those leading to slip and weakening.
Experiments modeling rupture along an entire fault were
performed for a variety of earthquake sizes and rupture speeds.
Each simulation produced normal stress perturbations on the
fault – in this case induced by rupture along a bimaterial interface
(a Weertman pulse) – which were then input into a granular
system. The granular system models the fault gouge found in the
fault core and was held under the same normal and shear stress
conditions as the ruptured fault. Normal stress perturbations on
WELCOME
2016 SCEC Annual Meeting page 221
the granular material were aimed at inducing slip in the grains,
and capturing the weakening mechanisms at play, such as pore
pressurization and acoustic fluidization.
The exact mechanisms leading to slip, or the changes that cause
slip, are still poorly understood. Developing an understanding of
this process relies on accurately modeling the fault-scale input
and the granular reaction. By linking together the granular
behavior with the larger scale fault dynamics we have the
opportunity to examine relationships that previously have not
been studied. Linking together two vastly different scales may be
the key to understanding the physics controlling earthquake
behavior.
Surface slip behavior of the southern and central Alpine
fault, New Zealand, Jozi K Pearson, and Nicolas C Barth
(Poster Presentation 114)
The Alpine Fault (AF) is a major continental plate boundary
between the Australian and Pacific plates, linking oppositely-
dipping subduction zones in New Zealand. Due to its relative
geometric simplicity, high slip rate, and quasi-regular earthquake
recurrence, the AF is a paradigm for characterizing “ideal” plate
boundary behavior. It offers a stepping stone to understanding
earthquake behavior on comparable but more complex systems
such as the San Andreas Fault (SAF) System in California. A
burgeoning record of more than ten paleoearthquakes on the AF
spanning over four hundred kilometers of fault length reveals
multi-mode earthquake behavior wherein some earthquakes
propagate across section boundaries (e.g.: representing
changes in geometry, slip vectors, and seismicity) and others
appear to arrest at these locations. One critical section boundary
is located between the southern and central AF sections.
Through geomorphic analysis of high resolution light detection
and ranging (lidar) data and supporting fieldwork on the southern
and central sections of the AF, we aim to supplement the
paleoseismic record with along-strike measurements of single
earthquake surface displacements to determine spatiotemporal
variations in slip accumulation and slip rate. We will also inform
seismic hazards (specifically the location and width of surface
rupture) and structural complexity, which may influence
earthquake rupture propagation or arrest. Offset features will be
dated by the following Quaternary dating methods:
dendrochronology, optically stimulated luminescence (OSL),
cosmogenic radionuclide (CRN), and/or Carbon-14 methods.
The primary hypothesis is that surface rupture offsets along the
AF will follow a uniform slip model in which repeating large
earthquakes have essentially the same slip distribution in each
event and that the total slip accumulation across the southern-
central AF section boundary will have low spatiotemporal
variation, which is consistent with previously determined slip
rates and paleoearthquakes. Understanding slip distribution
along the AF will clarify controls on continental plate boundary
behavior, thus providing grounds for new research directions on
other similar continental plate boundaries that exhibit more
complexities and uncertainties.
Performance enhancements and visualization for
RSQSim earthquake simulator, Dmitry Pekurovsky, Amit
Chourasia, Keith B Richards-Dinger, Bruce E Shaw, James H
Dieterich, and Yifeng Cui (Poster Presentation 347)
We report work in progress on RSQSim earthquake simulator
code (Dieterich/Richards-Dinger). One goal is to streamline and
optimize the algorithm to make it more scalable when running on
high core counts on modern supercomputer platforms, such as
Blue Waters (NCSA/UIUC), Stampede (TACC). Load imbalance
is the major bottleneck inherent in the algorithm. It creates
congestion and waste of resources by making some processors
idly wait until others finish their calculation. To overcome load
imbalance several algorithmic modifications are being
considered. By changing the order of computations it is possible
to reduce the amount of imbalanced compute load, and thus
improve overall performance.
In addition, visualization work is underway.
RSQSim generates a seismic event catalog for several thousand
years. This catalog includes rich detail about the quake which not
only includes the magnitude, location and time, but also which
patches got affected and the details associated. Effective study
of this catalog requires visualization to provide context in space
and time on where an event have occurred. To accomplish this
goal we have created web browser based interactive
visualization that shows a selected earthquake event on a map,
and also indicated preceding and upcoming event trails as well
as patches that were affected. This work involved creation of a
database from raw files, this database can be easily queried to
fetch necessary information on demand. The visualization
routine, uses that database on the backend to fetch data on
demand and plot it on the map, which is displayed on web
browser. Additionally we display 1000 year time series
visualization of earthquake events in the neighborhood of
selected events. The interface enables the scientist to
interactively explore the RSQSim seismic catalog.
Investigating Complex Slow Slip Evolution with High-
Resolution Tremor Catalogs and Numerical Simulations,
Yajun Peng, and Allan M Rubin (Poster Presentation 203)
Significant complexities of episodic slip and tremor (ETS) have
been revealed by short tremor bursts lasting minutes to hours,
many of which show clear migration patterns. In Cascadia, large-
scale rapid tremor reversals (RTRs) extend tens of km along
strike, repeatedly occupying the same general source area
during an ETS episode [e.g. Thomas et al, 2013; Peng and
Rubin, 2016]. We also observe repetitive tremor bursts occurring
well behind the main front in Guerrero, Mexico. In contrast to
RTRs, these bursts do not originate from the main front, and
generally propagate along the slip direction, similar to those
reported from Shikoku, Japan [Shelly et al., 2007]. Both types of
bursts occur intermittently, with recurrence intervals gradually
increasing to tidal periods. However, even the tidally-modulated
bursts are unlikely to be driven solely by tidal forcing. Since the
stress must decrease during each burst, while the local maxima
of the tidal stress remain nearly constant, each tidal peak stress
cannot supply the stress drop for the next repetition.
Here we explore the possibility that these repetitive bursts are
driven by surrounding tremor-less slow slip. We develop a
numerical model governed by a rate-and-state friction law that
transitions from velocity-weakening to velocity-strengthening
with increasing slip speed. A region with a larger transitional
velocity than the background is used to represent the tremor
zone. For this zone to slip intermittently, its stiffness needs to be
WELCOME
2016 SCEC Annual Meeting page 222
sufficiently large that the slip during each burst is less than the
total slip of the background during an episode, but smaller than
its own critical stiffness. This critical stiffness decreases as the
ratio of the background loading rate to the transitional cutoff
velocity increases; from elasticity this ratio decreases as the main
front moves across the model tremor zone. With these
considerations, we successfully reproduce the burst-like
behavior with increasingly large recurrence intervals in the model
tremor zone during a single slow slip event. Future work will
include investigating the propagation velocities of these bursts,
which in Guerrero decrease systematically with increasing time
since the previous migration through the same region, and tidal
modulation of their recurrence intervals.
Microseismic event detection using local coherence:
applications to the Long Beach 3D array and the Hi-
CLIMB linear array, Zhigang Peng, Zefeng Li, Dongdong Yao,
Daniel D Hollis, and William Frank (Poster Presentation 233)
We recently developed a new method for seismic event
detection, specifically designed for ultra-dense nodal-type
seismic arrays [Li and Peng, in prep]. This method involves a
function termed “local coherence” which quantifies the
coherence of signals between the examined station and its
nearest 5 neighbors. We apply this method to one-week
continuous data collected by the 5200-sensor Long Beach array,
and detected many very weak events that were buried in
background noise and not listed in the local earthquake catalogs.
To test its performance on a regular dense array, we also apply
it to the Hi-CLIMB array across the Himalayan frontal thrusts. The
phase I array of the project has 81 broadband stations, and the
inter-station distance is 3-5 km. During one-month test period, we
detected numerous distant, regional and local seismic events.
While most of them are regular earthquakes, some are likely
surface events such as landslides. We plan to apply the method
to the entire time period and classify the detected events into
different types. Updated results will be presented at the meeting.
Shallow level incipient pulverization found with a
restraining bend of the Clark segment of the San Jacinto
Fault, southern California, Daniel W Peppard, Heather N
Webb, and Gary H Girty (Poster Presentation 077)
This poster characterizes the properties of incipient pulverization
within a contractional bend along the Clark segment, San Jacinto
fault. About 1.5 km SE of the study site, the Clark segment
bifurcates into two branches. The NE-most splay, the topic of this
study, is located within a meander bend of Alkali Wash where the
Pleistocene Bautista Fm structural underlies metasedimentary
rocks of the Mesozoic Burnt Valley complex. Within this bend,
the fault lies at ~200 m below a Tertiary erosional surface, and is
marked by an ~30 cm thick black fault core that dips ~60° SW,
but further to the SE shallows to ~12-22° SW. The damage zone
within the Bautista Fm is subdivided into inner and outer zones.
The outer damage zone lies ~35-75 m below the over thrust
Burnt Valley complex. It consists of both undamaged and
damaged sandstones that give way toward the black fault core to
cataclasite series rocks within the ~10-22 cm wide inner damage
zone. In undamaged sandstones, matrix makes up ~17-21%
while framework minerals exhibit little if any evidence for in situ
cracking. In contrast, matrix within damaged sandstones makes
up 16-36%, and some, but not all, quartz framework grains are
fragmented. Within the inner damage zone, evidence for
cataclastic flow is ubiquitous, and matrix exceeds 71% in matrix-
rich banded cataclasites. In unbanded cataclasites and matrix-
poor banded cataclasites, matrix ranges from ~67-70%. The
results of clay mineralogical studies of the < 2 µm fractions
indicates that the matrix in undamaged sandstones and in matrix-
rich banded cataclasites consists of illite(I) + kaolinite(Ka) +
mixed-layer illite/smectite (I/S) with ~10% I. The clay mineral
assemblage in damaged sandstones is similar, but the %I in I/S
varies from ~10% to 20% I. In the unbanded cataclasite series,
the clay mineral assemblage is like that in the damaged
sandstones but the %I in I/S varies from ~30% to 40%. In matrix-
poor banded cataclasites, the clay-mineral assemblage includes
either I + Ka, or I + Ka + I/S with 20% I. The clay assemblage in
undamaged sandstones likely reflects the cool and wet
Pleistocene climate in which the Bautista Fm evolved. The
increase in I in I/S in unbanded and in matrix-poor banded
cataclasites likely reflects fluid-driven K metasomatism. The clay
mineralogy in the matrix-rich samples is similar to that in
undamaged sandstones, and suggests that fluid flow through the
damage zone was disconnected and temporally irregular.
New slip rates and characterization of active faults in the
northern Walker Lane, Ian Pierce, Steven G Wesnousky, and
Lewis A Owen (Poster Presentation 097)
Antelope, Mason and Smith Valleys are half-grabens within the
Northern Walker Lane, east and south of Carson City and Reno,
Nevada. We apply recently acquired 0.5-1 m/pixel lidar data and
terrestrial cosmogenic nuclide (TCN) surface exposure ages to
characterize the geometry, Quaternary expression, and slip rates
of active faults in these three valleys. The rangefront fault along
the western margin of Antelope Valley displays numerous
uplifted and faulted alluvial fans and streams, of which several
display apparent right-lateral deflections and/or offsets.
Cosmogenic CL-36 dating of a boulders on a fan that is vertically
offset by ~30 m place a maximum bound of ~0.6 m/yr on the
vertical displacement rate, about half of recent geodetic
estimates. The majority of the rangefront fault in Mason Valley
forms an alluvial-bedrock contact, and displays only one site
where the active fault has cut and displaced older fan sediments.
TCN ages of a ~10 m vertically faulted alluvial fan surface at this
site are >100 ka, suggesting a <0.1 mm/yr vertical slip rate for
the range bounding fault, significantly less than geodetic
estimates. Lidar also reveals a previously unmapped northeast
striking fault outboard of the rangefront at the northernmost part
of Mason Valley that offsets Holocene sediments and shows
morphology suggestive of left-lateral? strike-slip. Analysis of lidar
in Smith Valley shows a much more active range front fault than
that of Mason Valley, with frequently offset late Quaternary fan
surfaces along the northern half of the range, along with two
additional previously unmapped northwest striking fault
segments towards the center of the basin that also offset
Holocene sediments. The Artesia Fan along the Smith Valley
rangefront is offset ~10 m, and TCN ages range from 10-30 ka,
leading to a vertical slip rate of 0.33-1.0 mm/yr. The rate in this
case appears to be in line with geodetic rates of deformation that
have been reported by others.
CFM Version 5.1: New and revised 3D fault
representations and an improved database., Andreas
WELCOME
2016 SCEC Annual Meeting page 223
Plesch, Craig Nicholson, Christopher C Sorlien, John H Shaw,
and Egill Hauksson (Poster Presentation 003)
We present an updated version of the SCEC Community Fault
Model (CFM) for southern California. For version 5.1, we focused
on improving the database component of CFM which is critical
for the internal consistency and maintainability of the model. This
system also enables model consumers to access and assess the
full richness of the various modeled fault systems and their
alternative 3D fault representations. The improved database now
contains more than 800 rows of named fault-component objects
organized in a sensible hierarchical scheme. These components
are fault sections, splays and strands comprising faults and fault
systems. The database is now synchronized with the latest
catalog of individual t-surf CFM fault representations. This
synchronization included recent updates to the database with
over 300 new fault-component objects developed for the San
Andreas fault system and Eastern Transverse Ranges, Mojave
& Offshore Continental Borderland fault areas. Through this
process, we identified more than 30 fault components that
needed updating and were improved either by developing new
3D alternative fault models or by better regularization of their 3D
fault surface meshes. as well as by a more consistent
assignment to suitable locations in the hierarchical system. In
addition, important independent fault attributes such as average
strike and dip and the fault's equivalent USGS Quaternary Fault
(qfault) id association were updated to reflect any changes in
underlying fault geometry.
As a result, many new or updated 3D representations were
added to CFM in the Mojave and Eastern Transverse Ranges
fault areas, together with updated, revised models for the
Newport-Inglewood-Rose Canyon fault system and faults in the
Western Transverse Ranges associated with the Ventura
Special Fault Study Area. These fault updates in CFM include
extending the active North Channel-Pitas Point-Red Mountain
fault system west to Pt. Conception, updating the geometry of the
offshore Oak Ridge and Mid-Channel faults, and incorporating
sub-horizontal faults linked to S-dipping faults in the western,
eastern, and mid-channel areas, including the Hondo, Padre
Juan & Oak Ridge faults based on mapping with industry seismic
reflection data. Further, a number of these faults were extended
to greater depths with available relocated seismicity.
Trimming the UCERF3-TD Hazard Tree with a New
Probabilistic Model-Reduction Technique, Keith A Porter,
Edward H Field, and Kevin R Milner (Poster Presentation 300)
The Uniform California Earthquake Rupture Forecast Version 3,
Time-Dependent (UCERF3-TD) logic tree has 5,760 branches,
which can be computationally problematic for risk analysis of
large portfolios, especially when multiplied by five NGAWest-2
ground-motion prediction equations and two site-effect models.
An insurer or catastrophe risk modeler concerned with rare,
catastrophic losses to a portfolio of assets would have to
evaluate the portfolio 57,600 times to create a loss exceedance
curve that explores the entire possibility space. Which branches
matter most, and which can be ignored? We employed two
model-reduction techniques to find a subset of UCERF3-TD
parameters that must vary and fixed baseline values for the
remainder such that the reduced model produces approximately
the same distribution of loss that the full model does. The two
techniques are (1) the same tornado-diagram approach we
employed previously for UCERF2, and (2) an apparently novel
probabilistic sensitivity approach that is better suited to functions
of nominal random variables and that produced a smaller
reduced model with only 60 leaves, as opposed to 57,600.
Results can be used to reduce computational effort in loss
analyses by orders of magnitude.
Internal structure of the San Jacinto fault zone in the
trifurcation area southeast of Anza, California, from data
of dense linear arrays, Lei Qin, Yehuda Ben-Zion, Hongrui
Qiu, Pieter-Ewald Share, Zachary E Ross, and Frank L Vernon
(Poster Presentation 220)
We analyze two years of data from a linear array composed of
six 3-component sensors separated by about 25 m, and one-
month data from 1108 vertical geophones in ~650 m x 650 m box
configuration with instrument spacing 10-30 m, straddling the
Clark branch of the San Jacinto fault zone southeast of Anza,
California. The examined data include recordings from >10000
local events and some teleseismic events. Automatic picking
algorithms and visual inspection are used to identify P and S
body waves, along with fault zone head and trapped waves. The
locations of the seismogenic fault and damage zone are found
by examining waveform changes in both along-strike fault-
normal directions. The seismogenic fault location agrees with a
geologically mapped surface trace of the Clark branch. Regional
earthquake tomography shows that the NE side of the fault has
faster seismic velocity on large scale. Statistical analysis of P
arrival times indicates a local reversal of the velocity contrast
across the fault, with slower velocities NE of the fault related to
the fault damage zone. A portion of the damage zone generates
P trapped waves. Waveforms recorded by several fault-normal
lines of the box array are stacked to improve the ability to detect
fault zone head and S trapped waves. Updated results will be
presented in the meeting.
Internal structure of the San Jacinto fault zone at Jackass
Flat from data recorded by a dense linear array, Hongrui
Qiu, Yehuda Ben-Zion, Zachary E Ross, Pieter-Ewald Share,
and Frank L Vernon (Poster Presentation 219)
We analyze seismograms recorded by a dense linear array
crossing the San Jacinto Fault Zone (SW to NE) at Jackass Flat
(JF) southeast of the trifurcation area. The array has 9 stations
with ~20-30 m separation in the center and total aperture of ~400
m. Data of events recorded in 2012-2014 are used to search for
fault zone head and trapped waves. Delay times between phases
generated by ~3500 local earthquakes are utilized to estimate
velocity variations across the array. We use automatic P picks for
each source-receiver pair to calculate the ratio of each station
pick to the array-wide average pick. Statistical analysis of all data
indicates an initial increase followed by flattening and then
decrease of relative slowness from SW to NE. The results
suggest that all JF array stations are inside a fault damage zone.
We perform a systematic search for fault zone head waves using
an automatic detection algorithm and visual inspection. A set of
events shows a clear moveout between head and direct P waves
within the JF array with increasing moveout from SW to NE.
However, data generated by different events and observed at a
given station show no moveout with increasing source-receiver
distance, implying that the head waves are generated by a local
WELCOME
2016 SCEC Annual Meeting page 224
interface (perhaps the edge of the damage zone or basin) rather
than a deep fault interface. Fault zone trapped waves are
identified with automatic detection and visual inspection. The
generating events have broad spatial distribution, implying a
relatively shallow trapping structure. The trapped waves are
observed clearly only by a few stations on the SW part of the
array where the relative slowness is near maximum. Sets of
waveforms recorded across the entire array, with candidate-
trapped waves at the SW stations, are inverted with a genetic
algorithm for fault zone parameters. The most likely parameters
for the trapping structure are depth of 2-3 km, width of 100-200
m, Q value of 10-20, and velocity reduction of 40-50% from the
host rock. Combining the results with the surface geology, the
seismogenic fault is inferred to be located between the SW end
stations JFS3 and JFS4.
Evaluation of Ambient Noise Green’s Functions Using
Synthetic Ground Motions, Santiago Rabade, Leonardo
Ramirez-Guzman, Alan Juarez, and Jorge Aguirre (Poster
Presentation 234)
One of the recent topics of discussion in ambient noise
processing is the reliability and validation of the Green's
Functions obtained using cross-correlation techniques. We
compared the latter against synthetic seismograms computed for
reported velocity models. The data utilized for this comparison
and validation process was gathered from a dense array
deployed during 2014 in northern Mexico. We processed the
three-component 10Hz geophones array recordings for six hours
to obtain Green’s functions based on noise cross-correlation. On
the other hand, we built a 3D velocity model considering
reflectors and P- and S-wave velocities reported from the active
source and Spatial Autocorrelation experiments, respectively.
Using the model we computed synthetic seismograms with the
Carnegie Mellon Finite Element Toolchain, Hercules (Tu et al.,
2006). Preliminary results show agreement between the ambient
noise-based Green's functions and the synthetic seismograms in
different frequency ranges.
Biomarker thermal maturity at seismic timescales in high-
velocity rotary shear experiments, Hannah S Rabinowitz,
Heather M Savage, Elena Spagnuolo, and Giulio Di Toro (Poster
Presentation 026)
Studies of past seismicity on sampled fault structures often rely
on evidence of frictional heating on slip surfaces. Estimates of
temperature rise on a fault provide constraints on the size of
previous earthquakes hosted on the fault, allowing for a more
thorough understanding of the earthquake history of seismically
active regions of the world. One method that has been developed
to estimate temperature rise on faults is biomarker thermal
maturity, a tool that has previously been used to understand the
production of hydrocarbons through burial heating over millions
of years. The application of this method to paleoseismicity is
based on the fact that the organic material in sediments will
change when heated even over short timescales, as long as
temperature increase is large. This concept has been
demonstrated in multiple faults around the world (Pasagshak
Point Megathrust, AK; Punchbowl Fault, CA; Japan Trench) and
kinetics of thermal maturation have been established
experimentally for several biomarkers (methylphenanthrenes,
alkenones, n-alkanes). However, the experimental range to date
has been limited to times over ~20 min. While these short-
duration experiments provide evidence that biomarkers thermally
mature at time-scales significantly shorter than those relevant to
hydrocarbon formation, the kinetics still require extrapolation to
time-scales relevant to seismic slip. Here, we present high
velocity rotary shear experiments conducted on SHIVA (Slow to
HIgh Velocity Apparatus) at INGV in Rome. Samples of
Woodford shale were sheared at slip velocities of 1 m/s for a
range of displacements in order to test whether
methylphenanthrenes thermally mature at times between 1 s and
1 min, yielding temperatures up to ~300 ˚C (as measured from
thermal couples at the edge of the slipping zone). Thin sections
show that slip localizes to a ~100 µm zone in all experiments that
achieve earthquake speeds; one experiment run at 1 mm/s did
not show localization. We observe that methylphenanthrene
thermal maturity positively correlates with frictional work and
power density. We compare the experimental thermal maturity
change to that expected from extrapolation of the established
kinetics to the appropriate time/temperatures conditions. These
experiments provide confirmation that biomarker thermal
maturity can be used to map and quantify earthquake
temperature rise in fault zones.
Effects of realistic fault geometry on simulated ground
motions in the 2010 Darfield Earthquake, New Zealand,
Hoby N Razafindrakoto, Brendon A Bradley, and Robert W
Graves (Poster Presentation 284)
This study investigates the effects of earthquake source
information on simulated hybrid broadband ground motions,
using the 2010 Darfield Earthquake, New Zealand as a case
study. The geometric complexity of the fault for the 2010 Darfield
event provides an opportunity to analyse the sensitivity of
simulated ground motions to variations in the assumed fault
geometry and slip distribution. In the analysis, three different fault
geometries were considered, consisting of one, four and six fault
segments in which the overall seismic moments are the same.
Each of the aforementioned fault geometry cases consider ten
different slip distribution realizations.
Simulated ground motions resulting from each earthquake
source considered were computed using a hybrid approach,
combining low and high frequencies. The simulations were then
validated by comparing the simulated ground motions with
observed ground motion records. The residual statistical analysis
shows that the ground motion variability due to source geometry
and slip distribution is larger in the intermediate period range
(between 1 and 5s) and for stations near the source. The extent
of forward directivity also changes significantly between different
source models which is particularly observed at stations located
in the Christchurch Basin. Finally, the results reveal that a single
fault segment cannot appropriately represent the 2010 Darfield
event. The comparison of simulation results for one and four fault
segments, for example, shows no significant changes for the
stations located to the east of the causative faults, whereas
variability of the spectral acceleration residual appears along to
the west for intermediate periods (1s and 3s). This, among other
observations, implies that appropriate complexity is required,
particularly on the west end of the causative fault. The results of
this study have implications on the ground motions produced by
geometrically complex faults and the seismic hazard in the
Canterbury region.
WELCOME
2016 SCEC Annual Meeting page 225
Towards morphologic and cosmogenic dating of
paleoearthquake and fault slip rates in the Central
Nevada Seismic Belt, Tabor J Reedy, and Steven G
Wesnousky (Poster Presentation 104)
The Pleasant Valley Fault and Dixie Valley Fault are located
approximately 30km and 90km south of Winnemucca NV, within
the Central Nevada Seismic Belt a region of aligned, nearly
continuous, historical normal-fault earthquakes that produced
surface rupture. We have commenced study at two sites. The
first, in Pleasant Valley, is along a small ~1 km section of the
Pearce scarp near Siard Canyon that ruptured during the 1915
Ms 7.7 earthquake and the second is at Terrace Creek in Dixie
Valley that ruptured during the Ms 6.8 earthquake of 1954. At
each site, aerial photos collected from a drone are processed
with Agisoft PhotoScan software to produce high-resolution
DEMS. Preliminary analysis of scarp profiles with a linear scarp
diffusion model across the scarp near Siard Canyon suggest that
the 1915 earthquake was preceded ~14.5 ky ago by an event of
~ 4 m vertical displacement. Sediment samples taken from a 2 m
deep pit excavated on the hanging wall have been collected for
cosmogenic dating of the upper alluvial surface and will be used
as an internal consistency check on the scarp diffusion analysis
and place a bound on the late Pleistocene slip rate of the
Pleasant Valley scarp. At Terrace Creek the Dixie Valley fault
truncates a sequence of 3 well preserved fluvial terraces and
pediments which record a total of ~20 m of uplift, and 2 or more
higher and older pediment surfaces, all of which record a late
Pleistocene record of continued uplift. Sediments sampled from
a 2-m pit on the higher of the 3 well-preserved terraces is
expected to provide a bound on the late Pleistocene slip rate of
the fault.
Late Holocene rupture history of the Ash Hill fault, Eastern
California Shear Zone, Christine Regalla, Hannah Pangrcic,
Eric Kirby, and Eric McDonald (Poster Presentation 092)
Recent paleoseismic investigations in the Eastern California
Shear Zone (ECSZ) suggest that the Panamint, Garlock, and
Owens Valley faults may have experienced a late Holocene
cluster of seismicity, concurrent with an ongoing <~1.5 ka cluster
documented in the southern ECSZ. The spatial and temporal
patterns in seismicity are complex, however, and constraints on
earthquake recurrence along faults in the northern ECSZ are
limited. Here we present new mapping of alluvial deposits offset
by the Ash Hill fault, a ~50km long, near-vertical, right-lateral
transtensional fault in the western Panamint Valley, that provide
evidence for three ruptures since ~2-4 ka. We develop an alluvial
stratigraphy from field observations, air photos, and LiDAR
topography based on the relative development of desert varnish,
desert pavement, modification of bar and swale morphology, and
surface clast weathering. Single and multi-event fault scarps
within the youngest five generations of alluvium occur along >30
km of fault length. Vertical and lateral offsets of alluvium
measured at >80 locations define three apparent clusters of
displacement magnitude, with the largest displacements
restricted to the oldest units. At one site, a single continuous
surface rupture laterally displaces three generations of alluvium,
with 1-1.5m, 1.6-2.6m, and 2.7-4.9m of offset in the youngest,
intermediate, and oldest surfaces. Ages of faulted surfaces,
currently determined using a locally calibrated soil
chronosequence, suggest that the most recent rupture occurred
within the past ~500-800 years, and that the previous two
ruptures occurred since 2-4 ka. When combined with
displacement magnitudes, these data indicate late Holocene slip
rates on the order of 0.5–1 mm/yr, with a ~5:1 ratio of right lateral
to extensional slip, and ~0.9-2m of displacement per event. The
timing of events on the Ash Hill fault overlaps with age estimates
for late Holocene ruptures on the Panamint and Owens Valley
faults, and suggests the potential for strain transfer and
temporally clustered seismicity across both the southern and
northern ECSZ.
Role of fault geometry on the spatial distribution of the slip
budget, Phillip G Resor, Michele L Cooke, Scott T Marshall, and
Elizabeth H Madden (Poster Presentation 015)
A fundamental problem in earthquake physics is how stress is
transferred from plate motion to faults. Kinematic models assume
that long-term fault slip rates will sum to the plate velocity;
however, in a number of locations in southern California slip rates
determined from modeling geodetic data in this manner differ
significantly from geologic estimates. In this study we use
mechanical models to investigate how releasing steps may affect
estimates of fault slip over geologic time scales.
A suite of 2D models of idealized fault systems reveals how fault
length, friction and step geometry affect kinematic efficiency of
the system and the distribution of slip rate along the faults. We
find that although systems with longer fault segments are more
efficient, accommodating ~56-86% of plate displacement, point
sampling of these systems during geologic studies is unlikely
(p<50%) to yield representative slip rate estimates. Systems with
short segments, although much less efficient, are more likely to
yield representative slip rate estimates, particularly when slip on
overlapping segments is summed.
A 3D model of the San Jacinto fault illustrates how fault system
geometry may impact slip rate estimates along a real fault
system. Modeled slip rates are faster in the middle of fault
segments and slower within releasing steps, consistent with
published geologic estimates. The mean slip rate along the
modeled fault trace, however, is significantly slower than the
simple average of the geological rates. Our results suggest that
the location of geologic slip rate studies may govern their
suitability for hazard estimates and that models can be used to
put point measurements of slip into the context of slip distribution
throughout a fault system.
Evidence for Abrupt Subsidence Event in Carpinteria
Marsh at 2 ± 0.2 ka, Laura C Reynolds, Alexander R Simms,
Thomas K Rockwell, John M Bentz, and Robert B Peters (Poster
Presentation 087)
Recent work by Simms et al., 2016 suggests Carpinteria Marsh,
Carpinteria, California, has experienced an average of 1.2±0.4
mm/yr of subsidence in the last 7ka. However, the nature
(ongoing vs. abrupt) of this subsidence remained unclear. Here
we use a detailed stratigraphy of 43 vibracores and 7 Geoprobe
cores, along with 62 radiocarbon dates to show evidence for one
occurrence of abrupt subsidence at 4 ±0.5 m below MHHW. At
this depth, a sharp surface between organic-rich, fine-grained
deposits and overlying fine gray sand exists in all cores
penetrating below 4 m. The consistency of the elevation of the
surface and lack of erosional signatures indicate an abrupt
WELCOME
2016 SCEC Annual Meeting page 226
environmental change. The underlying fine-grained deposits
contain numerous organic fragments, Cerithidea californica
shells (a high-marsh species of gastropod), and low to mid marsh
foraminifera species (Jadammina spp., etc.). We interpret these
deposits to represent marsh surface elevations (MSL-MHHW).
The sand above the surface is devoid of foraminifera, but
includes shells of subtidal (below MSL) species such as Ostrea
spp., Macoma nasuta, and Saxidomus nuttallii. Radiocarbon
dates taken below and above this surface in 6 cores dates it to 2
± 0.2 ka. The abrupt deposition of subtidal deposits on supratidal
deposits requires an increase in relative sea level. Because no
dramatic fluctuations in sea level are known for this time period,
we attribute this drowned surface to rapid subsidence of the
marsh surface.
Earthquake likelihood models for New Zealand combining
information on strain rates, earthquakes and faults, David
A Rhoades, Annemarie Christophersen, and Matthew C
Gerstenberger (Poster Presentation 304)
We have constructed a set of multiplicative hybrid earthquake
likelihood models for the New Zealand CSEP testing region, in
which cell rates in a spatially uniform baseline model are scaled
using combinations of covariates derived from the earthquake
catalogue, fault data, and strain rate estimates. We considered
three components of the strain rate estimated from GPS data
over the period 1991-2011: the shear, rotational and dilatational
strain rates. Hybrid model parameters were fitted to target
earthquakes of M 5 and greater over the period 1987-2006 and
independently tested on earthquakes from the period 2012-2015.
According to information gain statistics, the shear strain rate is
the most informative individual covariate in the fitting and testing
periods, and this is confirmed by Molchan error diagrams. Most
models including strain rates are significantly more informative
than the best models that can be constructed without them. A
hybrid model that combines the shear and dilatational strain rates
with a smoothed seismicity covariate is the most informative
model in the fitting period, and a simpler model without the
dilatational strain rate is the most informative in the testing
period. These results can be used to improve earthquake source
modelling in probabilistic seismic hazard analysis. Also, strain
rates can usefully be incorporated into the background
component of short- and medium-term earthquake forecasting
models, such as ETAS and EEPAS.
Surface topography effects in three-dimensional physics-
based deterministic ground motion simulations in
southern California, Andrea C Riaño, Doriam Restrepo,
Ricardo Taborda, and Jacobo Bielak (Poster Presentation 294)
Topographic features are known to contribute to ground motion
amplification, especially at mountain ridges and edges. These
amplifications occur due to the constructive interference between
incoming motion with wavelengths comparable to the physical
dimensions of the surficial irregularities. Because of their nature,
these effects tend to be more influential at local scales, and
particularly significant at higher frequencies (above 1 Hz). In the
past, these effects have been typically ignored in most 3D low-
frequency regional-scale earthquake ground motion simulations.
However, simulations above 1 Hz are now commonplace, and a
variety of effects associated with higher frequencies need to be
considered, including topography. In this study we perform a
series of 5-Hz 3D simulations for the region of the Oxnard plain
including the surrounding geologic structures, and focus on the
effects of the topography on the ground response. We do so by
comparing the results from models with and without the surface
topography. Simulations are done using Hercules, a finite
element parallel code part of the SCEC High-F simulation
platform. Hercules implements a virtual topography method
which allows one to perform ground motion simulations explicitly
incorporating the effects of the surface topography in the models.
This method has been tested in various idealized and realistic
conditions, but it has not been used previously in simulations for
southern California. We study the Oxnard plain region because
of the presence of the Ventura Hills, San Cayetano Ridge, Oak
Ridge, Simi Hills and the Santa Monica Mountains, where
topographic amplification has been observed; and because this
is an area under significant seismic hazard, which poses
considerable risk for the growing communities of Ventura County.
Our simulations highlight the relevance of topographic effects for
the southern California region and contribute to advancing the
development of physics-based ground motion prediction. The
simulations were done on Blue Waters at the National Center for
Supercomputing Applications.
Open Intervals, Clusters and Supercycles: 1100 years of
Moment Release in the Southern San Andreas Fault
System: Are we Ready for the Century of Earthquakes?
Thomas K Rockwell (Oral Presentation 9/12/16 10:30)
Compilation of paleoseismic data from several dozen trench sites
in the southern San Andreas fault system, along with geomorphic
observations of displacement in recent earthquakes, allows for
sequencing of the past 1100 years of large earthquakes for the
southern 160 km of the main plate boundary. At least four
generalizations are clear: 1) M7 and larger earthquakes account
for most of the moment release in the southern San Andreas fault
system over the past 1100 years; 2) large earthquakes on
individual faults are quasi-periodic but display a relatively high
coefficient of variation in recurrence time, similar to most long
California paleoseismic records, and which may be a reflection
of Coulomb stress interactions; 3) moment release has
temporally varied during the past 1100 years but within
potentially predictable bounds; and 4) the southern San Andreas
fault system is currently moment deficit when compared to the
previous millennium of moment release. Comparison of the
timing of earthquakes with the timing of past high-stands of Lake
Cahuilla supports the idea that lake fillings influenced and may
have modulated the timing of southernmost San Andreas fault
earthquakes. If so, then the long (300 year) open interval on the
southern San Andreas fault may be in part explained by the
absence of a complete lake filling episode in the past 300 years.
Together, the record suggests that the southern San Andreas
fault is late in the cycle but not necessarily “overdue”, and that a
systems level approach may be more accurate in long-term
earthquake forecasting than estimates made from individual
faults. If the past is the key to the future, then the next century is
likely to see far more large earthquake activity on the southern
San Andreas fault system than was observed in the past one to
two centuries.
Strain accumulation on faults beneath Los Angeles: a
geodesy-based picture accounting for the effects of
WELCOME
2016 SCEC Annual Meeting page 227
sedimentary basins and anthropogenic surface
deformation, Chris Rollins, Donald F Argus, Walter Landry,
Sylvain D Barbot, and Jean-Philippe Avouac (Poster
Presentation 148)
The Los Angeles region is contracting at ~8 mm/yr in the N 5° E
direction due to the misalignment of the Mojave section of the
San Andreas Fault with the direction of relative Pacific-North
American plate motion. This contraction is accommodated by the
accumulation of strain on thrust faults such as the Sierra Madre,
Puente Hills and other systems and the release of that strain in
damaging earthquakes such as the 1971 San Fernando, 1987
Whittier Narrows and 1994 Northridge shocks. A larger
earthquake on one of these systems could constitute a worst-
case-scenario event for Los Angeles, and so it is essential to use
geodetic data to constrain where, and how quickly, tectonic strain
is accumulating on these faults. This estimation problem is
affected by 1) anthropogenic surface deformation that overprints
tectonic contraction in geodetic data, 2) the complex 3D
geometries of the relevant faults, and 3) the soft sedimentary
basin underlying Los Angeles, which affects the elastostatic
Green’s functions that map slip rates on faults to velocities at the
surface. Using 1) a GPS velocity field corrected for
anthropogenic motions [Argus et al, 2005, and in prep.], 2) a
detailed quadrilateral mesh of fault geometry based on an
updated version of that in Marshall et al [2009] and on the SCEC
CFM5.0 [Shaw et al, 2015], and 3) elastostatic Green’s functions
that incorporate the lateral and vertical heterogeneities in elastic
properties represented by the SCEC CVM-H15.1, we obtain the
most accurate geodesy-based picture of strain accumulation
beneath Los Angeles to date. Among other results, we find that
strain accumulation on strike-slip faults such as the Palos
Verdes, Whittier-Elsinore and Raymond-Hollywood systems may
cause an apparent N-S contractional gradient of ~2 mm/yr across
Los Angeles that is unrelated to thrust faulting, and that inferred
strain accumulation rates on thrust faults are more readily
reconciled with geologic slip rates when this strike-slip motion is
taken into account.
Complex active faults and seismicity rates in the
trifurcation area of the San Jacinto fault zone illuminated
by aftershocks of the 2016 Mw 5.2 Borrego Springs
earthquake, Zachary E Ross, Egill Hauksson, and Yehuda Ben-
Zion (Poster Presentation 218)
We summarize spatial properties and rates of early aftershocks
of the Mw 5.2, June 2016, Borrego Springs earthquake and
background seismicity in the trifurcation area of the San Jacinto
Fault Zone. In this area, the fault zone splits into three primary
sub-parallel strands and is associated with broad Vp/Vs
anomalies. A template matching technique is used to detect and
locate more than 25,000 early aftershocks that illuminate highly
complex active fault structures, in conjunction with high-
resolution regional seismicity. The hypocenters form lineations of
seismicity both along strike and normal to the main structure,
various dipping streaks of seismicity, coincident with interlaced
strike-slip and normal faults at depth. Events larger than M > 4
are located on the three main faults, while much of the small
magnitude aftershocks occurred in a damage zone between the
Clark and Buck Ridge strands, which spans several kilometers.
This spatial separation of events in different magnitude ranges
also is manifested in bimodal frequency-magnitude distributions
of events close to and away from the main faults. The results
provide important constraints on the physics of faulting at depth.
High-Frequency Nonlinear Simulations of Southern San
Andreas Earthquake Scenarios, Daniel Roten, Kim B Olsen,
Steven M Day, and Yifeng Cui (Poster Presentation 267)
The Southern San Andreas fault (SAF) is particularly likely to
host a large earthquake in southern California, with a 19%
chance of producing at least one M 6.7 earthquake during the
next 30 years (UCERF3). Large-scale computational efforts to
simulate large (M 7.5+) earthquakes on the SAF have shown that
predicted ground motions are sensitive to 3D basin effects,
dynamics of the seismic source (rupture speed, directivity) and
nonlinear behavior in the fault zone and on soft sedimentary fills.
However, as previous simulations of large San Andreas
earthquake scenarios were typically limited to frequencies of up
to 1 or 2 Hz, synthetic ground motions were only relevant for
structures vulnerable to long-period ground motions, in particular
high-rise buildings.
Here, we simulate M 7.7 earthquakes on the southern San
Andreas fault with a spatial resolution of 25 m, which allows us
to resolve frequencies up to 4 Hz using a minimum shear-wave
velocity of 500 m/s. We use a two-step method: First, dynamic
rupture on a planar, vertical fault roughly following the surface
trace of the SAF between Indio and Lake Hughes is simulated for
a compact computational domain using AWP-ODC CPU.
Second, seismic waves propagating from this source are
simulated within a larger box size including the densely
populated Los Angeles basin (LAB) using AWP-ODC GPU. Both
simulation steps are performed for a linear medium and a non-
linear medium governed by Drucker-Prager plasticity. Because
plasticity absorbs part of the seismic moment in the fault zone,
we use a fault boundary condition to impose near-fault particle
velocities obtained during dynamic simulation (step 1) in the
kinematic simulation (step 2). The mesh is defined by SCEC
CVM-S4, including a geotechnical layer, small-scale
heterogeneities and frequency-dependent quality factors.
Spectral accelerations at 2s (2s-SAs) frequently exceed 0.5g
near the fault and in the Los Angeles basin in the linear case.
Nonlinearity reduces long-period ground motions in these
locations by 25 to 50%. At higher frequencies, plastic effects are
most pronounced within ~20 km from the fault, where 1s-SAs are
reduced by up to 50% and 0.5s-SAs by up to 75% with respect
to linear simulations. However, accounting for plasticity brings
synthetic 0.5s-SAs in agreement with ground motion prediction
equations (GMPEs) at these fault distances, while linear 0.5s-
SAs exceed GMPEs by more than one standard deviation.
Effects of regulation on induced seismicity in southern
Kansas, Justin L Rubinstein, William L Ellsworth, and Sara L
Dougherty (Poster Presentation 181)
The appearance of seismicity concurrent with the expansion of
oil and gas activities in southern Kansas since September 2012
suggests that industrial operations are inducing earthquakes
there. Much of the seismicity can be related to high-rate injection
wells within 5 km of the earthquakes. There is significant
complexity to the situation, though. Some of the seismicity,
including the 2014 M4.8 Milan earthquake, the largest
WELCOME
2016 SCEC Annual Meeting page 228
earthquake to occur in the area, lies at least 10km from high-rate
injection wells. Additionally, the presence of high-rate wells does
not guarantee that there will be nearby seismicity. Many of the
highest-rate injection wells are located to the southwest of our
study area, where there is minimal seismicity.
We have also seen changes in earthquake rates shortly following
the March, 2015 enactment of new limits on the rate of
wastewater disposal in five areas in southern Kansas. Overall,
the earthquake rate has decreased significantly since these rules
went into place. In more detail, however, earthquake rates within
the five areas decreased, but the rate outside the five zones
increased. It is likely that fluid-pressure diffusion is responsible
for the migration of seismicity outside the areas of reduced
injection because there is little injection in the areas unaffected
by the new injection rules. This increase is also a reminder that
seismicity can persist long after the reduction or cessation of
injection. In addition to the effect of the new injection rules, it is
possible that the reduction in injection may be partially caused by
economic factors that have resulted in a decrease in the
production of oil and gas. We have yet to disentangle the effects
of the new injection rules and the low prices of oil and gas on the
induced seismicity in southern Kansas.
An Investigation into the Eastern Extent of the Garlock
Fault using Ground-based Magnetics and Seismotectonic
Analysis, Brad Ruddy, and Jascha Polet (Poster Presentation
245)
The character of the Garlock fault has been a topic of contention
among geologists since its discovery in 1910. The 250 km long
Garlock fault strikes roughly east-west and its current surface
expression is located between the San Andreas fault and the
Death Valley fault zone. Initially, it was believed that the Garlock
fault had only 8 km of displacement and that the eastern portion
of the Garlock turned south, turning into a westward dipping fault
underneath the Avawatz mountains. Later work done in the
1970’s determined that the displacement was much larger (48-
64 km) than previously suggested. This new data prompted the
hypothesis that the Garlock fault extends east of the Death Valley
fault zone. Magnetic and gravity measurements by Plescia and
Henyey (1982) across the extension of the fault suggested the
existence of an eastern limb of the fault, located in Kingston
wash. Their data showed contrasting magnetic basement
patterns: a gentle gradient to the north of the proposed fault
trace, with areas south of the fault exhibiting various highs and
lows, likely corresponding to buried horsts and grabens. The
purpose of our study is to confirm whether the Garlock fault does
continue east into Kingston wash, through the creation and
analysis of a walking magnetometer dataset at a much shorter
sampling distance. At present, four 8 km transects have been
surveyed with plans to collect 8-10 more over a 45 km2 study
area. This collected data confirms a large magnetic gradient in a
similar location as was established by Plescia and Henyey
(1982), which may represent the proposed eastern extension of
the Garlock fault. In addition to the magnetic data, 3-4 Very Low
Frequency Electromagnetics (VLF) walking transects will be
performed, in locations based on the preliminary analysis of the
magnetic data, to help create resistivity profiles. We plan to
develop a three-dimensional model to characterize the
eastwards extension of the fault at depth as well as the strike on
the surface. We also will generate maps and cross-sections of
the earthquake focal mechanism catalog by Yang et al. (2012)
and the relocated earthquake location catalog by Hauksson et al.
(2012) to assess whether any micro-seismicity corresponds to
the suggested strike and dip of the potential eastern extension of
the Garlock fault.
Fakequakes: Earthquake Rupture Scenarios for
California and Cascadia, Christine J Ruhl, Diego Melgar,
Ronni Grapenthin, Mario Aranha, and Richard M Allen (Poster
Presentation 319)
Limited geodetic observations of earthquakes greater than M6
make it difficult to test and improve GPS-based earthquake early
warning systems, such as G-larmS. Scenario ruptures and
ground motion simulations are therefore useful for studying
earthquake and tsunami hazards expected during future events.
We develop a catalog of scenario earthquakes on 25 large faults
(capable of M>6.5) in California, as well as the Cascadia
subduction zone, built from realistic 3D geometries. Synthetic
long-period 1Hz displacement waveforms are obtained from a
new stochastic kinematic slip distribution generation method.
Waveforms are validated by direct comparison to peak P-wave
displacement scaling laws and to PGD GMPEs obtained from
high-rate GPS observations of large events worldwide.
We will also discuss the generation of broadband synthetic
ground motions by hybrid deterministic/stochastic methods and
argue that they are still preferable, for these purposes, over fully
deterministic and computationally intensive forward modeling
approaches. Particularly, we will focus on portions of the
seismogram typically ignored in broadband synthesis: P-waves
and long period displacements that correctly capture static
offsets. It is important to adequately model these, in addition to
strong shaking, in order to study the response of currently
operating early warning systems.
Reducing Uncertainty in Ground Motion Prediction
Equations by Understanding Path Effects, Valerie J
Sahakian, Annemarie S Baltay, and Tom C Hanks (Poster
Presentation 175)
The differences between recorded ground motions and median
predictions made by Ground Motion Prediction Equations
(GMPEs) can be widely scattered, leading to large model
standard deviations which can result in large ground motion
amplitudes at low probabilities of exceedance. As such, reducing
uncertainty in GMPEs is integral in decreasing the variability in
predicted ground motions at low probabilities. Our approach for
reducing uncertainty in the predictions is to understand the
separate effects of source-, path-, and site-specific information.
The knowable, repeatable parts of these effects can thus be
identified and removed from populations of residuals, which
results in location-specific GMPEs. These terms can then be
associated with other geophysical information. This study
focuses on correlating the path term to material properties along
each recording’s raypath.
We employ a small database of approximately 3000 events
recorded on the Anza seismic network, as shown in Baltay et al.
(submitted), with magnitudes ranging from M 1 to M 5.4, with the
majority of the events in the range 1 < M < 3. The Anza network
has been in operation since ~1981, resulting in redundancy in
source-to-station paths. We show various models for a regional
WELCOME
2016 SCEC Annual Meeting page 229
GMPE, resulting from an inversion of all of the data. Next, we
decompose the residuals for the preferred model into source,
site, and path components, and focus on the path term. We
examine two tomographic models of the Southern California
region, the joint body wave and ambient noise model of Fang et
al. (2016) and Allam et al. (2014). We raytrace through the Fang
et al. (2016) model, and discuss various indices that can be
computed from both models representing the effects of seismic
attenuation, velocity structure on the raypath, and how they
correlate to the computed path term. Finally, we describe a plan
for moving forward with robust estimation of the effects of
material properties on the path term, and how to incorporate this
new path-specific knowledge into GMPE’s. We will discuss the
implications on the resulting reduction in uncertainty and hence
hazard level for low-probability earthquakes.
Topography effects in the near field for a dislocation point
source: A comparison of the FEM-HERCULES code and
the 3D IBEM, Edilson F Salazar Monroy, Leonardo Ramirez-
Guzman, Marcial Contreras-Zazueta, and Francisco J Sanchez-
Sesma (Poster Presentation 295)
We analyze a staircase and slope-dependant meshing strategy
for modeling topographical features using the HERCULES finite
element toolchain (Tu et al., 2006). We compare the Finite
Element solution against the Integral Boundary Element Results
using an ellipsoidal Gaussian shape as a benchmark and an
irregularity of the free boundary in an elastic half-space,
subjected to incident waves generated by a dislocation. The
staircase approach analyzed here is based on a slope-
dependent variable mesh which we tailored to follow the free
surface, naturally fulfilling the free boundary condition. The IBEM,
our reference solution, is based on an integral representation of
auxiliary force densities. This approach allows low-cost and
accurate computations for 3D configurations, as we only
discretized the free boundary and used Stokes' analytical
solution for the Green’s function of the full elastic space. Our
cross-verification uses ground motions computed at the free
surface using both numerical schemes. The results indicate that
elements with an average of 10 to 12 points per wavelength,
together with a few smaller ones, provide accurate solutions
when a low-pass filter is applied to remove high frequencies
associated with spurious eigenvalues of the discretized dynamic
system. Finally, we discuss advantages and limitations of these
numerical approaches.
Investigating the age and origin of small offsets at Van
Matre Ranch along the San Andreas Fault in the Carrizo
Plain, California, James B Salisbury, Ramon Arrowsmith,
Thomas K Rockwell, Sinan O Akciz, Nathan D Brown, and Lisa
Grant Ludwig (Poster Presentation 130)
Small displacement fault-offset features (<10’s of m) are rarely
dated, making it challenging to attribute slip to dated
earthquakes. We investigated the ages of subtle fluvial
depressions previously interpreted as beheaded channels
representing offsets (<10 m) at Van Matre Ranch (VMR) along
the San Andreas fault (SAF) in the Carrizo Plain. At VMR, several
catchments (~1,500-3,000 m2) NE of the SAF drain to the SW
and are truncated by the well-expressed SAF. We excavated four
fault-parallel trenches (T1-T4, NW to SE) across subtle
depressions ~10 m SW of the SAF. Single-grain feldspar pIR-
IRSL age estimates of channel fill indicate that only one channel
(T2) is young enough to be associated with a nearby feeder
drainage. The T2 sediments are offset 12 m and have a minimum
age estimate of 0.3±0.05 ka (pre-fading correction). The
deposition of these sediments (1666-1766 AD) falls within the
uncertainty range of the penultimate event at nearby Bidart fan
(1631-1863 AD, preferred 1713 AD). Hand excavated exposures
at T1 (35 m to the NW) suggest that T2 channel sediments have
experienced two earthquake events (1631-1863 AD and 1857
AD), as the third event back (at Bidart; 1580-1640 AD) predates
their deposition. At T1, hand-dug trenches show that the
“beheaded gully” is actually a fosse between two small offset
alluvial fans. Reconstruction of the young alluvial fan apex
suggests slip in 1857 was ~4 m. Therefore, slip in the penultimate
earthquake is estimated at ≤~8 m at the VMR site. We prefer this
interpretation but cannot discount that T2 channel sediments
were deposited when there was along-fault flow (before complete
beheading), thus making slip in the PE <8 m. These data suggest
a two-event, maximum slip rate of ~40 mm/yr – higher than the
30-37 mm/yr estimated from decadal geodesy and other slip rate
studies in the area. Interestingly, at trenches T4, T3, and T1,
buried channel ages are much older (3.8 ka, 4.7-5.3 ka, and 5.7-
7.3 ka, respectively). Coupling these results with a constant long-
term slip rate of 33-40 mm/yr suggests distant sources – larger
basins (~7,000-16,000 m2) beyond the SE end of our study area.
These interpretations indicate: a) there may be appreciable high-
frequency variation in paleoearthquake slip along strike; b)
topographic depressions have the potential to be, but are not
necessarily, reoccupied with continued slip; and c) small
catchments do not produce channel deposits as frequently as
has been assumed.
The SCEC Community Geodetic Model V1: Horizontal
Velocity Grid, David T Sandwell, Yuehua Zeng, Zheng-Kang
Shen, Brendan W Crowell, Jessica R Murray, Robert McCaffrey,
and Xiaohua Xu (Poster Presentation 141)
The SCEC community is constructing and updating a suite of
models for the Southern California region to facilitate cross-
disciplinary research (CFM, CVM, CGM, CSM, and CRM). Here
we are concerned with the development of the Community
Geodetic Model (CGM). Eventually the CGM will consist of vector
deformation time series at ~1 km resolution, and better than
seasonal sampling. As a first step we are constructing a 0.01˚
resolution grid of horizontal vector velocities and 2-D tensor
strain rate covering the areas of interest to SCEC scientists
(Figure 1). Our approach is to first assemble 15 available velocity
and strain rate models for the SCEC region. There were 4 main
approaches to model construction: isotropic interpolation,
interpolation guided by known faults, interpolation of a
rheologically-layered lithosphere, and model fitting using deep
dislocations in an elastic layer or a half space. We then evaluate
the 15 strain rate models in terms of roughness, cross
correlation, seismicity rate, and SHmax to select a subset of 9
usable models. Since all the models are based on slightly
different geodetic data and use a variety of reference frames, we
re-gridded velocities from the 9 models at a 0.01˚ grid spacing.
This is accomplished by forcing each velocity model to match the
best available GPS velocity data for the region. The 9 velocity
models were averaged (Figure 1) and their standard deviation
was also computed (Figure 2). Standard deviations are generally
WELCOME
2016 SCEC Annual Meeting page 230
small (< 0.5 mm/yr) in areas of good GPS coverage; areas of
large standard deviation illustrate where InSAR velocities will
contribute most. This uniform velocity is a first step in the
development of the full 3-D time dependent CGM. This result is
important for seismic hazard evaluation as well as InSAR time
series analysis. As new GPS and InSAR data become available
this SCEC community model will continue to evolve. The full
compilation of the GPS velocity data, the contributed models, and
the consensus products will be available on the SCEC web site.
Kinematic source generation based on fractal fault
geometries, William H Savran, and Kim B Olsen (Poster
Presentation 293)
Simulating seismic wave propagation at frequencies of
engineering interest (~10 Hz) requires a source description with
extraordinary complexity, such as produced by geometrically
rough faults. Although spontaneous ruptures on geometrically
rough-faults have been shown to produce realistic far-field
spectra and comparable fits with GMPEs, they are far too
computationally demanding for projects involving a large-scale
computational effort. Here, we present our implementation of a
kinematic rupture generator that attempts to capture the
processes of the rough-fault spontaneous rupture models, at
least in a statistical sense, that are responsible for generating
realistic source spectra to higher frequencies. Our method uses
sequential Gaussian co-simulation to simulate slip, peak-slip
velocity (PSV), and rupture velocity (vrup) based on two-point
statistics defined by a linear model of co-regionalization. We
incorporate 3 variogram basis functions at micro-scale (nugget),
medium scale (several hundred meters), and large scale (several
kilometers) to capture correlation effects at all scale lengths
between all source parameters and the initial friction. In addition,
we define a multivariate Gaussian model for one-point statistics
(μ, σ) that accounts for correlations between slip, PSV, and vrup.
We find strong correlations between (0.6-0.9) for all one-point
statistics except σvrup, which shows negative correlations of
approximately -0.5. This implies that ruptures that are
energetically more favorable (i.e. resulting in larger average slip)
tend to propagate at more similar vrup, and visa-versa. We
condition the one-point model on average slip to control the M0
of the event. With the one-point and two-point statistics of the
source parameters fully defined we parameterize an exponential
source-time function at each sub-fault. This fully describes the
spatiotemporal evolution of slip on the entire fault plane. Finally,
we use the local sub-fault geometry and material properties to
translate the slip-rate function into moment-rate tensor
components used for ground motion simulation.
Comparison of the rupture history of the southern San
Andreas fault with empirical data on fault displacement
and rupture length, Katherine M Scharer (Poster Presentation
121)
Offset geomorphic features provide estimates of slip used to
derive paleoearthquake moment magnitude based on empirical
relationships from historic ruptures (e.g., Wells and Coppersmith,
1994; Wesnousky, 2008). Obtaining these measurements is
difficult, as they are typically offset channels that must be serially
excavated and radiocarbon dated. Consequently, these
measurements are sparse. Along the well-studied 300-km-long
combined Mojave, Big Bend, and Carrizo sections of the
southern San Andreas fault, for example, only five locations have
such data. In this study, I combine data from these locations with
hundreds of undated lidar-based measurements of offset along
the same sections of the fault published by Zielke et al. (2012) to
develop a model of paleorupture length and displacement along
strike. To construct the model, I applied the 1200-year-long
rupture history in Scharer et al. (2016), which identifies the
minimum length of individual paleoruptures based on dated
paleoearthquakes at seven sites along the same sections of the
San Andreas fault. Each dated paleoearthquake is pinned to the
correlative, dated offset measurement and then extended along
strike through clusters of lidar-based (undated) offset
measurements. The ruptures are straightforward to model where
there are dated measurements or where lidar-based
displacements are well separated, such as along the southern
Big Bend section. Regions with denser lidar-based
displacements, such the central Carrizo section, are more
subjective to model, and will be highlighted here. I also present
alternative models that accommodate uncertainty in event
correlation due to radiocarbon dating limitations, but these do not
change the overall conclusions. In the preferred model, the
displacement curves have variable offset along strike and often
have one fairly short stretch with a peak displacement, visually
similar to the variability in historic ruptures on strike slip faults.
The longest permissible rupture occurred ca. 1550 A.D. and has
a 100 km section with over 5 m of displacement, similar to the
historic Fort Tejon rupture of 1857 (Zielke et al., 2012). Shorter
ruptures 75 to 100 km long have smaller average displacements,
less than 3m. This approach is robust in that it honors multiple
large datasets and up-weights better constrained data such as
dated offset measurements and high-quality lidar-based
measurements, while de-emphasizing lower quality lidar
measurements.
The Origin of High-angle Dip-slip Earthquakes at
Geothermal Fields in California, Martin Schoenball, Patricia
Martínez-Garzón, Andrew J Barbour, and Grzegorz Kwiatek
(Poster Presentation 183)
We examine the source mechanisms of earthquakes occurring
in three California geothermal fields: The Geysers, Salton Sea,
and Coso. We find source mechanisms ranging from strike slip
faulting, consistent with the tectonic settings, to dip slip with
unusually steep dip angles which are inconsistent with local
structures. For example, we identify a fault zone in the Salton
Sea Geothermal Field imaged using precisely-relocated
hypocenters with a dip angle of ~60° yet double-couple focal
mechanisms indicate higher-angle dip-slip on ≥75° dipping
planes. We observe considerable temporal variability in the
distribution of source mechanisms. For example, at the Salton
Sea we find that the number of high angle dip-slip events
increased after 1989, when net-extraction rates were highest.
There is a concurrent decline in strike-slip and strike-slip-normal
faulting, the mechanisms expected from regional tectonics.
These unusual focal mechanisms and their spatio-temporal
patterns are enigmatic in terms of our understanding of faulting
in geothermal regions. While near-vertical fault planes are
expected to slip in a strike-slip sense, and dip slip is expected to
occur on moderately dipping faults, we observe dip slip on near-
vertical fault planes. However, for plausible stress states and
WELCOME
2016 SCEC Annual Meeting page 231
accounting for geothermal production, the resolved fault planes
should be stable.
We systematically analyze the source mechanisms of these
earthquakes using full moment tensor inversion to understand
the constraints imposed by assuming a double-couple source.
Applied to The Geysers field, we find a significant reduction in
the number of high-angle dip-slip mechanisms using the full
moment tensor. The remaining mechanisms displaying high-
angle dip-slip could be consistent with faults accommodating
subsidence and compaction associated with volumetric strain
changes in the geothermal reservoir.
Regional evolution of network detection completeness in
Japan, Danijel Schorlemmer, Naoshi Hirata, Yuzo Ishigaki,
Kazuyoshi Z Nanjo, Hiroshi Tsuruoka, Thomas Beutin, and
Fabian Euchner (Poster Presentation 313)
An important characteristic of any seismic network is its detection
completeness, which should be considered a function of space
and time. Many researchers rely on robust estimates of detection
completeness, especially when investigating statistical
parameters of earthquake occurrence like earthquake rates.
Contrary to traditional approaches, we do not estimate
completeness using methods in which the completeness
magnitude is defined as the deviation of the frequency-
magnitude distribution from the linear Gutenberg-Richter
relation. Here, we present a method based on empirical data
only: phase data, station information, and the network-specific
attenuation relation. For each station of the network we estimate
a time-dependent distribution function describing the detection
capability depending on magnitude and distance to the
earthquake. For each point in time, maps of detection
probabilities for certain magnitudes or overall completeness
levels are compiled based on these distributions. Therefore, this
method allows for inspection of station performances and their
evolution as well as investigations on local detection probabilities
even in regions without seismic activity.
We present a full history of network detection completeness for
Japan and discuss details of this evolution, e.g. the effects of the
Tohoku-oki earthquake sequence. For practical purposes we
deliver completeness estimates for catalog data of selected
regions and document the conservative completeness estimates
researchers can use when investigating the JMA catalog in
different regions over different periods. All presented results are
published on the CompletenessWeb
(www.completenessweb.org) from which the user can download
completeness data from all investigated regions, software codes
for reproducing the results, and publication-ready and
customizable figures.
Fault-parallel shear fabric in the ductile crust of Southern
California imaged using receiver functions, Vera Schulte-
Pelkum, and Karl Mueller (Poster Presentation 023)
We present results on deep crustal deformation fabric from a new
receiver function method. The method uses the azimuthal
variation of P-to-S converted arrivals on the radial and transverse
component of receiver functions. Dipping isotropic velocity
contrasts and anisotropy contrasts with a plunging symmetry axis
generate receiver function arrivals with two polarity flips with
backazimuth. The azimuthal position of the polarity changes
relates directly to the strike of the dipping contrast or dipping
foliation. Unlike shear wave splitting, the converted arrival
method carries depth information since the delay time of the
conversion scales to the depth of the dipping or anisotropic
contrast. Also unlike shear wave splitting, the method only
requires contrasts involving weak (a few percent) Vp anisotropy
over a narrow (~2 km or more) zone of deformation to generate
a robust signal. Applying this new method to stations with data
available at the IRIS DMC in Southern California, we find strikes
of foliation or dipping contrasts that are strongly aligned with
surface fault traces. Some of the largest arrivals are seen
immediately adjacent to or beneath major, active faults. The
depths of the imaged structures extend well below the base of
the seismogenic crust, suggesting that fault-aligned shear fabric
is present in the lower crust and possibly in the uppermost
mantle. Our observations support deformation models that
predict faults root into wider shear zones continuing through the
ductile lower crust. Fault-parallel strikes are seen along the San
Andreas and other major strike slip fault branches, in the Eastern
California Shear zone, and across the Salton Trough. Differently
oriented strikes parallel thrust faults bordering the Transverse
Ranges and the Garlock fault. Weaker, similarly aligned fabric
appears to be pervasive in much of Southern California between
active faults in areas that are thought to be moving as a rigid
block (e.g., San Diego County). Possible future analysis using
the denser coverage provided by all three-component broadband
and short-period stations in the area with data available at
SCEDC as well as IRIS would help answer the following
questions: How narrow are the fault-related shear zones at depth
and do they continue into the upper mantle, what is their
geometry, and does the observed fabric match recent models of
deep crustal strain)? Is pervasive weaker fabric away from major
faults related to distributed deformation in the current transform
regime, or to fossil fabric in subducted Orocopia and Pelona
schists? The work could provide constraints for SCEC5’s
proposed efforts to develop a Community Rheological Model of
rock lithology and fabrics through Southern California.
Figure (see pdf): Preliminary map of receiver function mapped
strikes and depths using a subset of available networks. Bar color
shows depth, bar length is amplitude of the azimuthal degree-1
arrival in units of horizontal to vertical ratio. The largest amplitude
arrival at each station is shown. Black lines are fault surface
traces.
Challenges and Opportunities for Communicating about
Seismic Hazard, Timothy L Sellnow, Emina Herovic, and
Deanna Sellnow (Poster Presentation 324)
The complexity of seismic hazard information poses
communication challenges for scientists charged with publicly
communicating this information. Scientists must translate
complex information for nonscientists who, in turn, make risk-
based decisions about their property and welfare.
Misinterpretation in this communication process can move
beyond frustration for both scientist and their publics to crises
such as occurred in L’Aquila, Italy, in 2009. The L’Aquila case
and others have inspired a call for more research into the social
science of communicating about seismic hazard (Jordan, et al.,
2011).
WELCOME
2016 SCEC Annual Meeting page 232
This study addresses the call for communication research to
improve the understanding of seismic hazard by nonscientists.
Structured interviews were conducted with 21 attendees at the
annual meeting sponsored by the Southern California
Earthquake Center in Palm Springs, California, September 12-
16, 2015. Interviewees were selected based on their involvement
with public communication. Interviews were chosen as a means
of data collection to assure that participants were given the
flexibility needed to express their personal views on the subject.
For purposes of consistency, a structured set of questions based
on the best practices of risk and crisis communication were used
in all 21 interviews.
Transcripts of the interviews were studied to identify themes
surrounding the challenges and opportunities for communicating
seismic hazard. Inductive coding was used to identify emerging
themes. The authors read the posts several times "looking for
similarities, or patterns within the subsample" (Boyatzis, 1998, p.
46). Where possible, themes were combined into major themes
with related subthemes.
This poster summarizes six major themes: three challenges and
three opportunities. Themes focusing on the challenges for
improved communication included counter-messages from non-
scientific sources, audience discomfort with uncertainty and
probability, and maintaining interest during quiet periods where
earthquakes have not occurred recently. Themes focusing on
opportunities for improving communication included strategies
for consistently distinguishing between hazard and risk,
acknowledging and emphasizing that scientific data cannot and
should not always translate into actions for risk management,
and the emphasis on operational earthquake forecasting as
means for dispelling misinterpretations about earthquake
prediction. A model is provided that matches opportunities with
challenges and provides specific communication
recommendations is provided.
Using a fast similarity search algorithm to identify
repeating earthquake sequences, Nader Shakibay Senobari,
and Gareth J Funning (Poster Presentation 244)
Repeating earthquakes (REs) are the regular or semi-regular
failures of the same patch on a fault, producing near-identical
waveforms at a given station. Sequences of REs are commonly
interpreted as slip on small locked patches surrounded by large
areas of fault that are creeping (Nadeau and McEvilly, 1999).
Detecting them, therefore, places important constraints on the
extent of fault creep at depth. In addition, the magnitude and
recurrence interval of these RE sequences can be related to the
creep rate and used as constraints on slip models.
A pair of events can be identified as REs based on a high cross-
correlation coefficient (CCC) between their waveforms. Thus a
fundamental step in RE searches is calculating the CCC for all
event waveform pairs recorded at common stations. This
becomes computationally expensive for large data sets. To
expedite our search, we use a fast and accurate similarity search
algorithm developed by the computer science community (Yeh et
al., 2015; Zhu et al., 2016). Our initial tests on a data set including
~1500 waveforms suggest it is around 40 times faster than the
algorithm that we used previously (Shakibay Senobari and
Funning, SCEC Annual Meeting, 2014).
In this study we test our implementation of the algorithm by
searching for REs in northern California fault systems upon which
creep is suspected, but not well constrained, including the
Rodgers Creek, Maacama, Bartlett Springs, Concord-Green
Valley, West Napa and Greenville faults, targeting events
recorded at stations where the instrument was not changed for
10 years or more. We identify event pairs with CCC>0.85 and
cluster them based on their similarity. A second, location based
filter, based on the differential S–P times for each event pair at 5
or more stations, is used as an independent check. We consider
a cluster of events a RE sequence if the source location
separation distance for each pair is less than the estimated
circular size of the source (e.g. Chen et al., 2008); these are
gathered into an RE catalogue. In future, we plan to use this
information in combination with geodetic data to produce a robust
creep distribution model for all of the faults in this region.
Precursory Swarm and Earthquake Forecasting in Indo-
Nepal Himalaya, Prof Daya SHANKER, and Harihar Paudyal
(Poster Presentation 302)
We investigate seismicity data from 1963-2006 in the light of
three medium size earthquakes of 1980 (mb 6.1), 1984 (mb 5.6)
and 1999 (mb 6.6) occurred in the Western Nepal and its
adjoining Indian region which were preceded by well defined
patterns of precursory swarm. While scanning seismicity data
from 1999 to 2006, it was observed that two additional cases of
characteristic seismicity patterns in 1999-2006, and 2003-2006
carry on. In these two cases, though the anomalous seismicity
exists, no mainshock has occurred so far. After critical analysis
of the data, it is observed that the seismicity from 1999 onwards
fluctuates in the order as low-high-low phases. Extension of
Precursory Swarm followed by quiescence with a significant low
seismicity, which is still continuing, is a signal for the future
medium size earthquakes in the region.
Preliminary results on Late Quaternary slip rate of the
central Haiyuan Fault constrained by terrestrial in situ
cosmogenic nuclides dating, UAV and LiDAR surveys,
Yanxiu Shao, Jing Liu, Michael E Oskin, Jerome Van der Woerd,
Jinyu Zhang, Peng Wang, Pengtao Wang, Wang Wang, and
Wenqian Yao (Poster Presentation 089)
Kinematic parameters of active faults are essential for
understanding active tectonics, and evaluation of regional
tectonic and crustal deformation models. Many study focus on
large active faults and acquisition of their slip rates of multiple
time scales with advanced technology and methods. The
Haiyuan fault is a major left-lateral active fault in the northeastern
Tibetan Plateau, extending ~1000km from the Hala Lake to the
west to the Liupan Shan to the east. Its eastern section ruptured
to surface in 1920 with magnitude >=8. The slip rate of the
Haiyuan fault has attracted a lot of attention due to its important
role in transferring deformation from on-plateau to off-plateau.
Researchers have carried out slip rate determination along the
fault, but results are controversial. Gaudemer et al. (1995)
estimated that strike slip rate is ~ 11 ± 4 mm/yr since the
Holocene on the central segment. However, there was no reliable
geochronology dating at the time, they only assigned an
interglacial age as displacement accumulating age to estimate
rate. We revisited Gaudemer et al. (1995) study site, Honggeda,
for a more accurate slip rate of the central segment. We used
WELCOME
2016 SCEC Annual Meeting page 233
terrestrial LiDAR and small Unmanned Aerial Vehicle to survey
the offset terraces and their geomorphology, acquired DEM and
aerial imageries with ~0.1 m resolution. With such a fine
resolution, the subtle offset features on the active river bed are
clearly visible in the aerial imageries. These high resolution
topography data can help us to analyze terraces deformation and
subsequent degradation processes. Samples of terrestrial in situ
cosmogenic nuclides were collected from three terraces, which
would improve the slip rate estimates, together with topography
data.
Coulomb stress evolution over the past 200 years and
seismic hazard along the Xianshuihe fault zone of
Sichuan, China, Zhigang Shao, Jing Xu, Hongsheng Ma, and
Langping Zhang (Poster Presentation 020)
This study focuses on the M ≥ 6.7 earthquakes that have
occurred since 1816 on the Xianshuihe fault zone in southwest
China. The interseismic Coulomb stress accumulation and the
Coulomb stress changes caused by coseismic dislocation and
postseismic viscoelastic relaxation of the previous shocks were
computed for different periods on the relevant fault segments.
Based on these results, we analyzed the relationship between
time-adjacent strong shocks and the Coulomb stress evolution
before every earthquake. The analysis suggests that strong
earthquakes mostly occurred in the Coulomb stress
enhancement region caused by coseismic dislocation and
postseismic viscoelastic relaxation of the last earthquake.
Considering the Coulomb stress evolution at the fault planes of
the epicentral area before earthquakes, we found that the
Coulomb stress accumulation caused by the interseismic
tectonic loading was dominant for most strong earthquakes. For
some other earthquakes the stress changes caused by
coseismic dislocation and postseismic viscoelastic relaxation of
surrounding earthquakes were very significant, which may be
equivalent to the effect of interseismic tectonic loading lasting
hundreds of years. Based on the time-dependent probabilistic
risk model and the Dieterich (1994) model, we estimate the
background seismic activity and the future earthquake probability
for different fault segments, using long term seismic activity and
strong earthquake recurrence cycles. It is shown that the Bamei,
Selaha, and Kangding segments of the Xianshuihe fault zone
have high earthquake probability, and are likely to have strong
earthquakes. If energy is accumulated up to the year 2050, the
magnitude of an event on these three segments could reach Mw
7.2, Mw 7.0, and Mw 7.1 respectively, while if the S7 and S8
cascades rupture, the event on these segments could reach a
magnitude of up to Mw 7.2.
High-resolution imaging of the San Andreas Fault around
San Gorgonio Pass using fault zone head waves and
double-difference tomography, with implications for large
earthquake ruptures, Pieter-Ewald Share, Yehuda Ben-Zion,
Clifford H Thurber, Haijiang Zhang, and Hao Guo (Poster
Presentation 120)
We attempt to clarify the seismic velocity structure within and
around the complex San Gorgonio Pass (SGP) “structural knot”
of the San Andreas Fault (SAF) using fault zone headwaves
(FZHW) and a new double-difference tomography code
incorporating both event pairs and station pairs. The tomography
provides information on the velocity structure at scales >1 km,
while the FZHW give direct evidence for the existence and
continuity of bimaterial fault interfaces. Arrival times of P and S
waves generated by 9260 M>1 local events occurring from
1/1/2010 to 30/6/2015 within a 150 by 100 km region centered
on the SGP and recorded by 143 stations are inverted for P and
S velocity structures. Low velocity anomalies are observed, in
particular at depths <6 km, SE of Cajon Pass between the SAF
and San Jacinto Fault to the south, and at the intersection of the
SAF/Mission Creek Fault and Pinto Mountain Fault to the north.
Clear FZHW are found to be generated by 41 M>1.5 events
within a 20 km zone around the SAF, and to propagate along two
separate bilateral interfaces. The first extends from slightly NW
of Cajon Pass to the SGP region and is associated with a slower
seismic velocity on the SW side of the SAF. The second interface
extends from slightly NE of SGP, ends in the Coachella Valley,
and has an opposite velocity contrast. The low velocity
tomographic anomalies correlate well with the inferred
alternating slow sides of the SAF identified by the FZHW
analysis. The observed FZHW, opposite velocity contrasts, and
expected behavior of bimaterial ruptures have important
implications for large earthquakes on the SAF in the region. The
velocity contrast to the SE of SGP suggests that earthquakes on
that section of the SAF tend to propagate to the SE. The opposite
velocity contrast to the NW of SGP suggests that earthquakes on
that section tend to propagate to the NW. The results also
suggest that ruptures that enter the SGP area from either
direction would likely be arrested by the reversal of the velocity
contrast they would encounter with continuing propagation.
Exploring the suitability of the Quake-Catcher Network in
the USGS ShakeMap: Case studies from aftershocks of
the 2010-2011 Darfield and Christchurch, New Zealand
earthquakes, Liam J Shaughnessy, Danielle F Sumy, Elizabeth
S Cochran, Corrie J Neighbors, and Robert-Michael de Groot
(Poster Presentation 231)
Following the September 3, 2010, MW 7.1 Darfield, New Zealand
earthquake, the Quake-Catcher Network (QCN;
quakecatcher.net) team rapidly deployed 192 low-cost seismic
sensors around Christchurch. These CodeMercenaries
JoyWarrior JWF14F8 14-bit three-component Class C
accelerometers monitored and recorded aftershocks of the
Darfield earthquake, which included the February 21, 2011 MW
6.3 Christchurch event. In terms of performance, QCN sensor
tests suggest these data are potentially useful for ground-motion
studies, which could include ShakeMap [Evans et al., 2014].
Previously, the Mexican Red Atrapa Sismos network team
illustrated that ground-motion results from a dense array of QCN
sensors in Mexico can be used to produce accurate intensity
ShakeMaps for a MW5.9 and MW7.2 earthquake, respectively
[Dominguez et al., 2015]. Therefore, the purpose of the present
study is to conduct a feasibility test to show if aftershock data
collected by QCN sensors deployed in the epicentral region of
the 2010-2011 Darfield and Christchurch earthquakes are
suitable for integration into the USGS ShakeMap software and
produce useful results. We compare the residuals between our
ground-motion results with distance as compared to the Bradley
[2013] ground-motion prediction equation for both QCN and
traditional GeoNet strong-motion data, and find that the QCN
accelerometers provide on-scale data suitable for use in
ShakeMap. We examine ShakeMaps composed of QCN and
WELCOME
2016 SCEC Annual Meeting page 234
GeoNet data individually and combined, and investigate the
ground-motion residuals between the maps for the sake of
comparison. Cochran et al. [2011] found that peak ground
velocity and acceleration data and pseudo-spectral accelerations
(PSA) measured by QCN are comparable to those measured by
the GeoNet stations, thus we expect that the individual QCN and
GeoNet ShakeMaps will look very similar. In addition, Kaiser et
al. [2011] concluded that the dense QCN array in Christchurch
provides comparable estimates of ground-motion and site
response as the nearby GeoNet stations, and that QCN sensors
can be useful in microzonation studies. We therefore conclude
that QCN accelerometers provide a cost-efficient method to
obtain accurate ground motion data for use in ShakeMaps. In
areas of low station density, where traditional networks are
limited, QCN sensors can increase local station density and
provide higher-resolution mapping of ground-motion across
urban areas that may help to inform earthquake response.
Precise relative stress drops, Bruce E Shaw, William L
Ellsworth, Nana Yoshimitsu, Yihe Huang, and Gregory C Beroza
(Poster Presentation 193)
Earthquake stress drops are generally observed to be
independent of seismic moment but vary by three orders of
magnitude or more for any moment value. If correct, this predicts
far greater variability in high-frequency ground motions than is
observed. Because stress drop depends on the cube of the
corner frequency, the question has been raised if stress drop
variability is a source effect or measurement error. To address
this problem, we present a new method that gives precise relative
stress drop values for pairs and populations of earthquakes.
Based on spectral ratios for nearby events, this approach uses
the low- and high-frequency asymptotic values of spectral ratios
and avoids the need to estimate corner frequencies. By avoiding
the spectral shape near the corner frequency and attendant
source complexity, the method produces precise stress drop
ratios for pairs of events. The method is only weakly dependent
on the assumed high-frequency asymptotic behavior of the
spectrum. Closure rules for connected sets of ratios enable
quantitative measures of method uncertainties and extension of
the measurements from pairs to populations of earthquakes.
Validation results comparing stress drop ratios with more
traditional corner frequency based methods will be presented.
Further results showing underlying intrinsic variability in stress
drops, a question relevant to ground motion hazard estimates,
will be presented.
Testing and Reconciling Stress Drop and Attenuation
Models in Southern California, Peter M Shearer, Rachel E
Abercrombie, and Daniel T Trugman (Poster Presentation 222)
Earthquake stress drop is a fundamental source parameter,
implicit in many SCEC science goals. It is relatively easy to
estimate from seismic data, but hard to measure reliably and
well. The large uncertainties and scatter in results affect strong
ground motion prediction and limit our understanding of the
physics of the earthquake rupture process. In addition, improved
stress drop estimates will help in determining if there are any
systematic differences in rupture properties between natural and
induced earthquakes, and among background seismicity,
swarms, foreshocks, and aftershocks. To improve the quality and
reliability of stress drop measurements in Southern California, we
compare in detail two different approaches that analyze P-wave
spectra. Our goal is to investigate sources of consistency and
discrepancies, and to better quantify uncertainties in stress drop
estimates. P-wave spectra also contain information about
attenuation and other path effects that can be compared to
existing models of attenuation in southern California. Our longer-
term aim is to develop an improved approach to invert jointly for
more accurate and reliable stress drops, together with regional
attenuation and site effects.
Shearer et al. [2006] developed a large-scale, regional approach
involving stacking and averaging P-wave spectra to obtain
parameters for large catalogs of events. Abercrombie [2013,
2014, 2015a] developed a smaller-scale, more-detailed
approach in an attempt to obtain the best possible results for a
small number of the best-recorded earthquakes. To compare
these methods, we focus on two test regions with dense
seismicity, one near the Landers earthquake epicenter and one
around the Cajon Pass borehole, in which previous results
predict spatial variability in both earthquake stress drops and
attenuation. Here we present preliminary results for the Landers
region, which contains over 1000 aftershocks of the 1992
mainshock. Of these, we focus on several hundred earthquakes
for which we were able to obtain P-wave corner frequency and
stress drop estimates using both methods.
Eventually we plan to: (1) determine precise (relative)
parameters for the best-recorded events, (2) quantify
uncertainties in the measurements, (3) update the Shearer et al.
[2006] study to include post-2001 events throughout southern
California, and (4) investigate azimuthal variations in the derived
path-correction functions to determine if they are consistent with
the Hauksson and Shearer [2006] attenuation model for southern
California.
A new strategy for earthquake focal mechanisms using
waveform-correlation-derived relative polarities and
cluster analysis: Application to a fluid-driven earthquake
swarm, David R Shelly, Jeanne L Hardebeck, William L
Ellsworth, and David P Hill (Poster Presentation 217)
In microseismicity analyses, reliable focal mechanisms can
typically be obtained for only a small subset of located events.
We address this limitation here, presenting a framework for
determining robust focal mechanisms for large populations of
very small events. To achieve this, we resolve relative P- and S-
wave polarities between pairs of waveforms using their signed
correlation coefficients - a byproduct of previously performed
precise earthquake relocation. We then use cluster analysis to
group events with similar patterns of polarities across the
network. Finally, we apply a standard mechanism inversion to the
grouped data, using either catalog or correlation-derived P-wave
polarity datasets. This approach has great potential for
enhancing analyses of spatially concentrated microseismicity
such as earthquake swarms, mainshock-aftershock sequences,
and industrial reservoir stimulation or injection-induced seismic
sequences.
To demonstrate its utility, we apply this technique to the 2014
Long Valley Caldera earthquake swarm, which has been
interpreted have been initiated and sustained by an evolving fluid
pressure transient. In our analysis, 85% of events (7212 out of
8494 located by Shelly et al. [2016]) fall within five well-
WELCOME
2016 SCEC Annual Meeting page 235
constrained focal mechanism clusters, more than 12 times the
number with network-determined mechanisms. Of the
earthquakes we characterize, 3023 (42%) of have magnitudes
smaller than 0.0. We find that mechanism variations are strongly
associated with corresponding variations in hypocentral
structure, yet mechanism heterogeneity also occurs where it
cannot be resolved by hypocentral patterns, often confined to
small-magnitude events. Small (5-20º) rotations between
mechanism orientations and earthquake location trends persist
when we apply recent 3D velocity models. Although this
discrepancy might still be a velocity-model artifact, it could also
be explained by a geometry of en echelon, interlinked shear and
dilational faulting.
Green’s functions retrieved by Multi-Component C3 for
Ground Motion Prediction, Yixiao Sheng, Gregory C Beroza,
and Marine A Denolle (Poster Presentation 274)
We follow the Virtual Earthquake Approach, proposed by Denolle
et al (2013), to predict strong ground motion, based on the
surface response determined from the ambient seismic field.
Instead of just using cross-correlation (C1), we use C3, cross-
correlation of the coda of the cross-correlation, to estimate the
impulse responses. It has been shown that C3 can mitigate the
uneven excitation of the ambient seismic wavefield, and result in
more symmetric Green’s functions (Stehly et al, 2008 and
Froment et al, 2011). In addition to using vertical-vertical
component in C1, we utilize other components, such as vertical-
east and vertical-north to construct C3 from multiple components.
To keep retain amplitude information, we follow Denolle et al
(2013), using deconvolution interferometry to calculate C1. C3
amplitudes are then calibrated with the C1 amplitudes. We test
this approach on the 2016 Mw 5.2 Borrego Springs earthquake,
using station YN.TR02, 2.5 km away from the epicenter, as a
virtual source. We show that the Green’s function estimated by
multi-component C3, though noisier, are comparable to and more
symmetric than those estimated by C1. We observe large energy
on R-T, T-R, Z-T and T-Z components of the Green’s function
tensors, indicating for strong Love-Rayleigh conversions. Such
wave conversions can have a strong effect on the ground motion
prediction results.
Inferring Southern California Crustal Viscosity Structure
from the SCEC Community Velocity Model, William
Shinevar, Mark Behn, Greg H Hirth, and Oliver Jagoutz (Poster
Presentation 339)
In this study we constrain the viscosity of the lower crust through
a joint analysis of the role of rock composition on seismic P-wave
(Vp), S-wave (Vs) velocities and rheology. Previous research has
demonstrated robust relationships between seismic velocity and
crustal composition, as well as between the composition and
viscosity of the lower crust. Here we extend these analyses to
determine whether seismic velocity can be used as a robust
indicator of crustal viscosity. Our analysis follows a three-step
approach. First, we use the Gibbs free energy minimization
routine Perple_X to calculate equilibrium mineral assemblages
for a global compilation of crustal rocks at various pressures and
temperatures. Second, we use the Huet et al. (JGR, 2014) mixing
model and single-phase flow laws for major crust-forming
minerals to calculate bulk viscosity for the predicted equilibrium
mineral assemblages. Third, we use Perple_X to calculate
seismic velocities for the mineral assemblages. Our results are
strongly influenced by the choice of the garnet flow law, which is
poorly constrained at crustal conditions. However, assuming a
strong garnet flow law (with a strength similar to that of pyroxene)
we find a strong correlation between crustal viscosity and the
seismic Vp/Vs ratio. This implies that even in regions where
seismic velocity does not give a strong constraint on composition,
it still provides robust estimates of crustal viscosity. We apply our
method to seismic data from Southern California to calculate a
viscosity for the regional lower crust.
Surface uplift and time-dependent seismic hazard due to
fluid-injection: Cast studies from texas, Manoochehr
Shirzaei (Poster Presentation 163)
Increasing seismicity in the central USA since 2009 coincides in
space and time with wastewater injection. However,
observations of the surface deformation and physical models to
constrain the extent to which fluid migrates and to unequivocally
link the seismicity and wastewater injection are scarce. Here we
present examples from Texas and Kansas and show that
wastewater injection causes uplift at a few mm/year, detectable
using radar interferometric data. Then the evolution of crustal
strain and pore pressure is constrained using the measured uplift
and reported injection data through a poroelastic model. We infer
that > 1 MPa increase in pore pressure in rocks with low
compressibility triggers earthquakes including the Mw4.8, 17
May 2012 event, the largest earthquake recorded in east Texas.
Observations that only deeper wells are associated with
earthquakes, whereas large pressure change in more shallow
aquifers are not, highlights the importance of hydrogeology. The
frequency and magnitude of earthquakes increased, even while
the injection rates declined, owing to diffusion of pore pressure
from earlier periods with higher injection rates. We show that
surface deformation data are useful to evaluate the evolution of
pore pressure and earthquake potential in the vicinity of injection
sites.
Earthquakes on Compressional Inversion Structures -
Problems in Mechanics and in Hazard Assessment,
Richard H Sibson (Oral Presentation 9/11/16 18:00)
The dip-distribution of near-pure reverse-slip ruptures includes a
dominant ‘Andersonian’ cluster (dips of 30±5°) flanked by groups
of low-angle thrusts (dips of 10±5°), and moderate-steep reverse
faults (dips of 50±5°). These last are attributable in part to
‘compressional inversion’ – reactivation during crustal
contraction of normal faults inherited from earlier crustal
extension – identified by a distinctive structural-stratigraphic
signature and shown by seismic profiling to be widespread. It is
easier, in terms of required stress levels, to impose brittle
deformation on the crust during extension; compressional
inversion is thus an inevitable consequence of the Wilson Cycle
of oceanic opening and closure. Examples of seismically active
inversion provinces include NE Honshu, Japan, and the NW
South Island, NZ, with damaging inland earthquakes < M7.5.
Reactivation mechanics does much to explain the observed dip
distribution for reverse-slip ruptures, suggesting first, that low-
displacement faults are characterized by ‘Byerlee’ friction (μs ~
0.6), and second, that high fluid overpressures are needed for
continued reactivation of moderate-steep reverse slip faults.
Support for the latter comes from the existence of hydrothermal
WELCOME
2016 SCEC Annual Meeting page 236
vein systems formed by ‘fault-valve’ action around reverse faults
exhumed from seismogenic depths, and from geophysical
anomalies (bright-spot reflectors, anomalously high Vp/Vs, high
electrical conductivity) in the mid-crust of areas undergoing
inversion. Compressional inversion earthquakes are
predominantly ‘fluid-driven’ (H2O, CO2?) because failure on
such structures is likely induced by accumulating fluid
overpressure rather than by increasing differential stress.
Evaluating hazard from inversion structures is problematic
because their surface expression is often structurally complex
with misleading dip-separations, and may be obscured near the
margins of sedimentary basins. Complexity also arises from
competition between inversion structures and younger, more
optimally oriented thrusts. Inversion structures are likely
widespread throughout North America within the west coast plate
boundary zone (e.g. Ventura and Los Angeles Basins within the
Transverse Ranges), within mid-continental rift zones, and along
the Atlantic seaboard where inherited Mesozoic rift structures are
occasionally reactivated in reverse-slip earthquakes (e.g. 1982
Miramachi, New Brunswick; 1983 Goodnow, NY).
The SCEC Broadband Platform: Open-Source Software
for Strong Ground Motion Simulation and Validation, Fabio
Silva, Christine A Goulet, Philip J Maechling, Scott Callaghan,
and Thomas H Jordan (Poster Presentation 002)
The Southern California Earthquake Center (SCEC) Broadband
Platform (BBP) is a carefully integrated collection of open-source
scientific software programs that can simulate broadband (0-100
Hz) ground motions for earthquakes at regional scales. The BBP
can run earthquake rupture and wave propagation modeling
software to simulate ground motions for well-observed historical
earthquakes and to quantify how well the simulated broadband
seismograms match the observed seismograms. The BBP can
also run simulations for hypothetical earthquakes. In this case,
users input an earthquake location and magnitude description, a
list of station locations, and a 1D velocity model for the region of
interest, and the BBP software then calculates ground motions
for the specified stations.
The BBP scientific software modules implement kinematic
rupture generation, low- and high-frequency seismogram
synthesis using wave propagation through 1D layered velocity
structures, several ground motion intensity measure calculations,
and various ground motion goodness-of-fit tools. These modules
are integrated into a software system that provides user-defined,
repeatable, calculation of ground-motion seismograms, using
multiple alternative ground motion simulation methods, and
software utilities to generate tables, plots, and maps. The BBP
has been developed over the last five years in a collaborative
project involving geoscientists, earthquake engineers, graduate
students, and SCEC scientific software developers.
The SCEC BBP software released in 2016 can be compiled and
run on recent Linux and Mac OS X systems with GNU compilers.
It includes five simulation methods, seven simulation regions
covering California, Japan, and Eastern North America, and the
ability to compare simulation results against empirical ground
motion models (aka GMPEs). The latest version includes
updated ground motion simulation methods, a suite of new
validation metrics and a simplified command line user interface.
Natural polish in granitic rocks, Shalev Siman-Tov, Emily E
Brodsky, and Greg M Stock (Poster Presentation 054)
Fault mirrors are highly smooth and reflective rock surfaces that
are found in many shear zones around the world. Recent studies
suggest that fault mirrors are formed during high velocity slip on
faults and therefore may serve as an indicator for seismic slip. In
contrast, other studies suggest that fault mirrors may form under
high normal stress at sub-seismic velocities and at room
temperature. Fault mirrors are observed within the fault core of
many rock type environments including limestone, dolomite,
chert and rhyolite. However, to the best of our knowledge, they
are missing in faults hosted in granite. Moreover, mirror-like
surfaces form during high velocity rotary shear experiments in
many types of rock but not in sheared granite blocks.
The absence of fault mirrors in granite is surprising, particularly
since there exists extensive glacial polish on granitic bedrock.
Glacial polish describes the smooth and reflective rock surfaces
formed at the base of glaciers that carved the underlying
bedrock. In addition to their import for studies of glacial dynamics
and geomorphology, glacially polished surfaces may hold some
significance for fault mechanics. Glacial polish and fault mirrors
share many similarities. At field exposures they both present
highly smooth surfaces and striations that clearly point in the slip
direction. Studies on carbonate fault mirrors showed that
individual highly reflective surfaces are composed of a thin
nanograin layer. Preliminary SEM observations on samples
collected from granitic rocks at Yosemite National Park suggest
that these polished surfaces are also coated by an ultrathin
cohesive layer composed of nanograins.
Although there are clear differences between glacial and fault-
zone environments, the similarity between these textures, and
the fact that both are formed during shear, suggest that a similar
mechanism is responsible for their formation. The comparison
raises questions about the importance of high fluid contents and
extremely mobilized gouge in forming polish.
Marine terraces, isostasy, earthquakes, and uplift rates,
Alexander R Simms, Helene Rouby, Kurt Lambeck, and Angela
Roman (Poster Presentation 083)
Marine terraces line the southern California coast and have been
used for decades to constrain rates of tectonic uplift. However,
despite years of study two fundamental questions remain
concerning their relationship to tectonic uplift. First, how accurate
are previous estimates of tectonic uplift that ignore the impacts
of glacio-isostatic adjustment (GIA)? Second, is that uplift slow
and steady or punctuated by abrupt (co-seismic) events. In this
study, we adjust uplift rates across the Pacific Coast of California
for the impacts of GIA. We also present results from a shallow-
marine seismic survey within a coastal lake inset into a marine
terrace in Santa Barbara. We find that ignoring GIA results in
uplift rates that are on average 40% too high across the California
Coast. Furthermore, accounting for GIA brings uplift rates
calculated from marine terraces of different ages into better
agreement. Our seismic survey also reveals a seismic pattern
consistent with abrupt uplift of a marine terrace near Santa
Barbara. This data suggests that uplift of the marine terrace was
punctuated by 4-6 co-seismic (?) events.
WELCOME
2016 SCEC Annual Meeting page 237
Dehydration-Induced Porosity Waves and Episodic
Tremor and Slip, Robert Skarbek, and Alan W Rempel (Poster
Presentation 071)
Episodic tremor and slip (ETS) at plate interfaces takes place
where there is abundant evidence for elevated, near-lithostatic
pore pressures. In subduction zones and on the San Andreas
fault, tectonic tremor occurs at depths where there is evidence
for a mixture of viscous and elastic deformation along the plate
interface. Particularly in subduction zones, ETS occurs at
sufficiently great depths where chemical dehydration reactions
must act as the dominant fluid source. Here, we simulate fluid
and heat flow while tracking the location of a vertically oriented,
one-dimensional column of material as it subducts through the
slow slip and tremor zone. The material in the column is
transformed through a pressure-dependent and temperature-
dependent dehydration reaction that we describe with a
generalized nonlinear kinetic rate law. Column deformation is
largely dominated by viscous creep, with a closure rate that
depends linearly on porosity. This behavior causes the
dehydration reaction to generate traveling porosity waves that
transport increased fluid pressures within the slow slip region. To
explore the possibility that the observed periodicity of slow slip
and tremor in subduction zones can be explained by the
migration of such porosity waves, we derive a dispersion relation
that accurately describes our numerical results. We also obtain
an expression for how the thickness of the dehydrating layer is
expected to vary as a function of the parameters in the reaction
rate law. Although the amplitudes of pore pressure perturbations
rival those that are produced by known external forcings (e.g.,
tides or passing surface waves), our analysis suggests that given
reasonable estimates of rock viscosity, permeabilities in the
range 6.5x10-15 to 5x10-10 m2 are required for porosity wave
trains to form at periods comparable to those of slow slip and
tremor.
Proximity of Precambrian basement affects the likelihood
of induced seismicity in the Appalachian, Illinois, and
Williston basins, Rob Skoumal, Michael Brudzinski, and Brian
Currie (Poster Presentation 189)
While most of the induced seismicity observed in the Central and
Eastern U.S. has occurred in sedimentary basins that have
experienced overall increases in oil and gas development (e.g.,
the Anadarko and Ft. Worth basins), other basins with similar
activity (e.g., the Williston and Northern Appalachian basins)
have experienced very little, if any, induced seismicity. While
hydro-geomechanical modeling indicates that induced seismicity
may be related to the proximity of critically stressed faults in the
crystalline basement, recent studies have found fluid injection
rate to be the dominant factor controlling induced seismicity. To
test these interpretations, we evaluated water disposal and well
completion records from the Appalachian, Illinois, and Williston
basins, and compared them with induced seismic sequences
identified through seismic template matching of all cataloged
earthquakes in these regions. Our results indicate a strong
correspondence between induced seismic events and the
proximity of subsurface wastewater injection/hydraulic fracturing
targets to crystalline basement rocks. For example, in the
Northern Appalachian Basin, of the >20 identified induced
seismic sequences, all but two were associated with
injection/completion targets located at depths within ~1 km of the
basement. In parts of the basin where target intervals are at
depths >1 km from basement, induced events have been
recorded only in proximity to basement-involved faults. In the
Williston Basin, most disposal interval/hydraulic fracturing targets
are >1 km above the crystalline basement which may explain the
lack of induced seismic events in the region despite high rate fluid
injection. Collectively, the results of our investigation suggest that
proximity to basement is an important variable in considering the
likelihood of induced seismicity associated with wastewater
disposal and hydraulic fracturing.
Frictional rheology for nonlinear attenuation: Implications
for paleoseismology and strong S-waves, Norman H Sleep
(Poster Presentation 279)
Strong Love waves and near-field velocity pulses impinge on
sedimentary basins and hard rocks in California. Rock damage
from past shaking provides paleoseismological evidence as well
as evidence that frictional failure caused the damage. The
dynamic strain eps is the ratio of particle velocity V to phase
velocity c. The dynamic stress tau_D is strain times the shear
modulus G, tau_D = G*V/c. Frictional failure occurs when tau_F
> mu*(P_lith-P_fluid)=~mu*g*z*(rho-rho_f), where mu is the
coefficient of friction, P_lith is lithostatic pressure, P_fluid is fluid
pressure, g is the accleration of gravity, z is depth, rho is rock
density, and rho_f is fluid density. The approximate equality
assumes that the water table is at the surface and constant
material properties. A testable hypothesis is that the shallow
subsurface self-organizes by damage in strong events so the the
dynamic stress just reaches the failure stress. Then, G*eps =
~mu*g*z*(rho-rho_f), so that G increases linearly with depth
within the damaged zone and G/z is constant. G/z does
predictably vary from a constant if mu varies with depth and if the
water table is not at the surface. Clay-rich rocks are expected to
have lower coefficients of friction than clay poor rocks. Borehole
with detailed well logs in sedimentary rocks near Parkfield and
accumulating sediments in the Santa Clara Valley show the
expected behavior. These results confirm that strong low-
frequency shaking sometimes occurs near Parkfield. Numerical
models of likely earthquakes are needed to obtain peak ground
velocity from strain. Frictional rheology should be used to model
high-frequency S-waves. The resolved horizontal acceleration
then clips near the effective coefficient of friction and the
horiozntal signal is transiently circularly polarized.
4D stress evolution models of the San Andreas Fault
System using improved geodetic and paleoseismic
constraints, Bridget R Smith-Konter, Karen M Luttrell, Xiaopeng
Tong, and David T Sandwell (Poster Presentation 010)
Major earthquakes of the San Andreas Fault System (SAFS) are
thought to occur when accumulated fault stress in the upper
locked portion of the crust exceeds some threshold value. 4D
simulations of stress evolution provide rare insight into
earthquake cycle crustal stress variations at seismogenic depths
where earthquake ruptures nucleate, however we emphasize
two important details: 1) modeled stress accumulation through
time is largely prescribed by the assumed slip history, and thus
these results highlight the need for continual improvements to
and utilization of contemporary paleoseismic chronologies; 2)
improved resolution and accuracy of the near-fault velocity field
using new, integrated GPS and InSAR data is critical for
WELCOME
2016 SCEC Annual Meeting page 238
improving strain and stress rate models of the SAFS, particularly
in areas where the GPS array spacing is inadequate for resolving
small variations in locking depth. Utilizing new SCEC community
paleoseismic and geodetic data, our refined stress models are
used to explore the temporally and spatially varying stress
threshold conditions, and the sensitivity of these results, for major
segments of the SAFS. 4D models demonstrate complex stress
accumulation and stress drop behaviors as functions of space
and time, with pockets of stress concentrated at depth due to the
interaction of neighboring fault segments and at fault segment
branching junctions. Because stress drops of major strike-slip
earthquakes rarely exceed 10 MPa, these models may provide a
lower bound on estimates of stress evolution throughout the
historical era, and perhaps an upper bound on the expected
recurrence interval of a particular fault segment. Moreover, these
4D visualizations lay the groundwork for 4D time-series
animations, an important step forward in simulating stress
evolution over multiple earthquake cycles spanning paleoseismic
timescales.
Magnitude and Peak Amplitude Relationship for
Microseismicity Induced by Hydraulic Fracture
Experiment, Trevor Smith, Adam Arce, and Chen Ji (Poster
Presentation 258)
Waveform cross-correlation technique is widely used to improve
the detection of small magnitude events induced by hydraulic
fracturing. However, when events are detected, assigning a
reliable magnitude is a challenging task, especially considering
their small signal amplitude and high background noise during
injections. In this study, we adopt the Match & Locate algorithm
(M&L, Zhang and Wen, 2015) to analyze seven hours of
continuous seismic observations from a hydraulic fracturing
experiment in Central California. The site of the stimulated region
is only 300-400m away from a 16-receiver vertical-borehole array
which spans 230 m. The sampling rate is 4000 Hz. Both the
injection sites and borehole array are more than 1.7 km below
the surface. This dataset has previously been studied by an
industry group, producing a catalog of 1134 events with moment
magnitudes (Mw) ranging from -3.1 to -0.9. In this study, we
select 202 events from this catalog with high signal to noise ratios
to use as templates. Our M&L analysis produces a new catalog
that contains 2119 events, which is ~10 times more detections
than the number of templates and about two times the original
catalog. Using these two catalogs, we investigate the relationship
of moment magnitude difference (ΔMW) and local magnitude
difference (ΔML) between the detected event and corresponding
template event. ΔML is computed using the peak amplitude ratio
between the detected and template event for each channel. Our
analysis yields an empirical relationship of ΔMW=0.64-0.65ΔML
with R2 of 0.99. The coefficient of ~2/3 suggests that the
information of the event’s corner frequency could be lost (Hanks
and Boore, 1984). The cause might not be unique, which implies
that Earth’s attenuation at this depth range (>1.7 km) is
significant; or the 4000 Hz sampling rate is not sufficient. This
relationship is crucial to estimate the b-value of the
microseismicity induced by hydraulic fracture experiments. The
analysis using the M&L catalog with the relation ΔMW = 2/3ΔML
results in a normal b-value of 1.1, while a smaller b value of 0.88
would be obtained if ΔMW = ΔML instead.
Stochastic Representations of Seismic Anisotropy:
Locally Isotropic Transversely Isotropic Media, Xin Song,
and Thomas H Jordan (Poster Presentation 236)
A self-consistent theory for the effective elastic parameters of
stochastic media with small-scale 3D heterogeneities has been
developed using a 2nd-order Born approximation to the scattered
wavefield [Jordan,2015]. Our model assumes the medium can be
represented as a spatially homogeneous, transversely isotropic
random field with a covariance tensor that can be factored into a
one-point, tensor-valued variance and a two-point, scalar-valued
correlation function. In the low-frequency limit, where the seismic
wavenumbers are small compared to the characteristic
wavenumbers of the heterogeneity, the locally isotropic
stochastic medium can be replaced by a homogeneous “effective
medium” with a transversely isotropic (TI) stiffness tensor that
depends only the one-point covariance matrix of the two Lamé
parameters and a dimensionless number η that measures the
horizontal-to-vertical aspect ratio of heterogeneity. If η = 1, the
heterogeneity is geometrically isotropic. The scattering caused
by a ~ 10% variation in seismic velocities reduces the effective
velocities by about 1%. As η --> 0, the medium is stretched into
a vertical stochastic bundle, and as η --> ∞, the medium is
flattened into a horizontal stochastic laminate. In the latter limit,
the expressions for the anisotropic effective moduli reduce to
Backus’s (1962) second-order expressions for a 1D stochastic
laminate. Comparisons with the exact Backus theory show that
the second-order approximation predicts the effective anisotropy
for non-Gaussian media fairly well with an error less than 10%
for media with velocity fluctuations less than about 15%, which
should be adequate for most crustal studies. We also note that
the 2nd-order theory is exact for heterogeneities that are gamma-
distributed. We apply the theory to heterogeneities in the Los
Angeles basin determined from well-log analysis [e.g. Plesch et
al., 2014] and compare the predicted anisotropy with seismic
constraints [Chen et al., 2007]. The observed shear wave
splitting ratios overlaps the values calculate from effective
medium theory with the relative Vp variations about 4-7% and η
in the range of 5 to 200.
Displacement direction and 3D geometry for the south-
directed North Channel – Pitas Point fault system and
north-directed ramps, decollements, and other faults
beneath Santa Barbara Channel., Christopher C Sorlien,
Craig Nicholson, Richard J Behl, and Marc J Kamerling (Poster
Presentation 007)
We use abundant seismic reflection data to interpret 3D fault
geometry in the upper 5 to 8 km beneath Santa Barbara Channel
and use dated and correlated seismic stratigraphy to study 3D
fold geometry and its changes through Quaternary time.
A set of north-directed (top-to-north) faults underlie much of
Santa Barbara Channel. The offshore S-dipping offshore Oak
Ridge fault merges westward with a blind fault forming a
kinematically-continuous 100+ km-long system. Horizontal N-
directed decollements intersect this broader Oak Ridge fault
system (ORFS) at 6 to 7 km depth. The down-to-north tilting
southern margin of Santa Barbara basin may be a forelimb above
the ORFS, instead of, or in addition to, being a backlimb above
the Channel Island thrust.
WELCOME
2016 SCEC Annual Meeting page 239
North Channel–Pitas Point fault strands dip north with S-directed
motion across 120 km from near Ventura to west of Point
Conception. This fault system is mostly blind and associated with
broad forelimb tilting. The rate of tilting decreases from east to
west but is relatively uniform in any one location since 1.8 Ma.
The geometry is complex. The 60 km-long offshore Pitas Point
fault strand continues farther east onshore as the Ventura fault.
To the west, 10 km west of UCSB, this NNE-dipping strand
terminates into N-dipping strands, forming a 25° bend, the Gato
Canyon segment boundary. Although N-dipping fault strands
continue west of this segment, the lower, southern faults die out
offshore Gaviota to be replaced by the north-directed Hondo
fault, linked southward by a subhorizontal decollement to a the
western ORFS. Blind N-dipping fault strands associated with a
tilting forelimb re-appear west of Gaviota and continue 30 km
farther to west of Pt. Conception. This segment may be
kinematically linked to a southern strand of the right-lateral
Hosgri fault.
Folding between the Gato Canyon segment boundary and Pitas
Point is partly accomplished by the Pitas Point strand together
with a stack of flatter N-dipping faults and the large S-dipping
Padre Juan (Hobson) blind fault. The Padre Juan fault is
interpreted for 30 km west of Pitas Point. Strands of the N-dipping
Red Mountain fault are also mapped 50 km between Carpinteria
and the Gato Canyon segment boundary. The continuity of the
progressive tilting above the western 50 km of the Pitas Point
fault suggests continuity of blind (oblique) thrust displacement. It
exhibits no evidence for a sea floor fault rupture for at least the
last half million years in that part.
Modern earthquake ruptures on contrasting fault systems
in the Laguna Salada region of northwestern Mexico.,
Ronald M Spelz, John M Fletcher, Thomas K Rockwell, Alejandro
Hinojosa, Michael E Oskin, Lewis A Owen, Jaziel F Cambron,
Víctor H Villaverde, Laura V Vallin, Abel A Gutierrez, and Keene
W Karlsson (Poster Presentation 073)
The Laguna Salada region of northwestern Mexico is located
near the axis of the Pacific-North American plate margin and
hosts a wide variety of active faults that accommodate the three
dimensional strain of transtensional shearing. Individual fault
segments can be classified by their slip tendency, which is high
for optimally oriented faults that form angles of ~30° from the
maximum compressive stress and ranges to very low values for
severely misoriented faults such as low angle normal faults. We
recognize three fundamentally different types of fault systems
that have contrasting mechanical properties, and we are
undertaking studies to determine if they also have contrasting
seismogenesis as reflected by their paleoseismic records of large
earthquake production during the past ~30 kyr. In 2010 the Mw
7.2 El Mayor-Cucapah earthquake activated a complex fault
system composed of individual faults that have a wide range of
orientations, extend short distances along strike (<30 km), and
typically end at intersections with other faults. Because each fault
is linked through multiple intersections with other faults, it is likely
that none of them cuts the entire seismogenic upper crust and
failure of the entire network thus should depend on a misoriented
fault that requires the largest differential stress (Fletcher et al.,
2016, Nature Geoscience). In 1892, the Mw7.2 Laguna Salada
earthquake activated the Laguna Salada fault, which controls the
northern half of the Laguna Salada basin and extends at least 75
km along strike, and thus is a simple fault system that must cut
the entire seismogenic upper crust. Along most of its length, the
Laguna Salada fault is optimally oriented and should slip under
relatively low magnitudes of applied differential stress. This
contrasts markedly with the Cañada David detachment, which is
the master fault that controls rifting in the southern half of the
Laguna Salada basin. With a strike length of ~55 km, this fault
also likely cuts the entire seismogenic crust, but its uniformly
shallow dip (<20°) at the surface shows that it is severely
misoriented. Therefore, it should require much higher
magnitudes of differential stress in order to slip. Systematic
mapping of Quaternary scarps and fan surfaces along the CDD
led to the discovery of the surface rupture produced by a ML =
6.5 earthquake in 1934. Although its focal mechanism is
characterized by subvertical nodal planes with pure strike slip
kinematics, we can show that the event actually originated on the
CDD. This discovery raises the worldwide total number of
modern earthquakes on low-angle normal faults to only nine
events and thus it provides an important opportunity to explain
the dichotomy posed by the paucity of seismologic evidence
versus the abundance of geologic evidence for large
earthquakes on low angle normal faults.
In summary, simple fault systems are those that cut the entire
seismogenic crust and we distinguish between the optimally
oriented Laguna Salada fault from the severely misoriented
Cañada David detachment. Complex fault systems, like the fault
array of the Sierra Cucapah, are composed of individual faults
that span a wide range of orientation-controlled slip tendency.
Individual faults have multiple intersections with others and
failure of the network is controlled by misoriented keystone faults
(Fletcher 2016, Nature Geoscience). Each of these three
contrasting fault systems has produced a modern earthquake
rupture and hosts a rich record of older earthquakes during the
past ~30 kyrs, which will illuminate potentially significant
differences in their mechanics and seismogenesis.
Body wave phases from higher order cross-correlation,
Zack Spica, and Gregory C Beroza (Poster Presentation 186)
Recent studies in ambient field seismology showed that body
wave phases (direct and reflected) can be extracted from
ambient-field cross-correlations. The latter offers the possibility
to sample Earth’s deep reflector (i.e. the outer core) in remote
regions of the globe where traditional imaging techniques lack of
earthquake signals. However, body wave phases retrieval using
seismic interferometry depends on synchronousness of the
seismic stations.
We show that we can go one step further and retrieve body wave
phase between asynchronous seismic network using spatially
stacked higher order cross-correlations (the so called C3). We
present results for California and Mexico and we show that both
direct P and/or ScS phases are clearly retrieved in the C3.
Extraction of deep phases by higher order cross-correlations can
therefore further improve the sampling of the deep Earth because
the technique can be applied anywhere broadband seismic
stations were once installed. In that sense, California is a good
target for this study because hundreds of asynchronous network
were installed since decades and we can take advantage of all
of them to improve the deep Earth model of the region.
WELCOME
2016 SCEC Annual Meeting page 240
Reevaluation of high slip rates on the Eglington Fault Las
Vegas, NV utilizing new chronostratigraphic and geologic
evidence, Kathleen B Springer, and Jeffrey S Pigati (Poster
Presentation 098)
The Eglington fault is a northeast striking monoclinal landform
that offsets middle-late Pleistocene groundwater discharge
deposits in the Las Vegas Valley by ~14 m vertically. Most of the
exposure of the fault has been lost to extensive urbanization, but
the scarp is preserved along its northeastern extent. The 2014
USGS National Seismic Hazards Map characterizes the
Eglington fault as a Quaternary fault and estimates the slip rate
at 0.16 mm/yr. Recent investigations indicate a significantly
higher slip rate (0.25 to 0.9 mm/yr, with a preferred rate of 0.6
mm/year). These estimates are based on pre-AMS radiocarbon
dates that constrain the age range of a deposit identified as “Unit
D” of the Las Vegas Formation (40-18 ka). The preferred offset
age of the fault was obtained from deposits that formed on the
upper part of “Unit D” (22 ka). Our recent work has redefined the
stratigraphy and chronology of the Las Vegas Formation in and
around the Eglington fault area, allowing reinterpretation of its
slip rate. Specifically, the Eglington fault does not displace the
entirety of “Unit D”, but only the middle bed (our Bed D2) from
which we obtained radiocarbon dates of 27.58-31.68 ka. Thus, a
more accurate revised slip rate is ~0.44 - 0.50 mm/year.
Additionally, we show that the 22 ka date that previously defined
the minimum displacement age for “Unit D” across the Eglington
scarp applies not to that unit but to a newly recognized unit
(Member E, Bed E0) that so far, has not been shown to be offset
by the fault. Additional investigation is necessary to determine
offset of this unit and other younger deposits in the sequence.
While calculations based on our revised stratigraphy and high-
precision chronology broadly corroborate previous high
contemporary slip-rate estimates for the Eglington fault, they also
significantly increase the precision of earlier slip rate estimates.
The revised rates also affect probable earthquake-recurrence
intervals and reinforce the need for further study of this fault with
respect to public safety in the Las Vegas Valley.
Estimates of fault tractions in the San Gorgonio Pass
region consider fault interaction, Aviel R Stern, Michele L
Cooke, Jennifer L Beyer, Jennifer M Tarnowski, David D
Oglesby, and Roby Douilly (Poster Presentation 058)
In order to estimate absolute tractions along the southern San
Andreas fault, we simulate the regional deformation using three-
dimensional quasi-static mechanical models. These models
simulate the stressing rates both over multiple earthquake cycles
and during interseismic loading of the fault system using deep
slip rates determined from the multi-cycle model. Assuming
complete shear stress drop during large ground rupturing events,
we estimate absolute shear tractions along faults using the time
since last event in the paleoseismic record for each fault segment
and the distribution of stressing rates from the interseismic
model. For the fault normal tractions, normal stressing rates
along the San Gorgonio Pass region locally accrue or reduce
after multiple earthquake cycles due to the complexity of the fault
geometry. The distribution of normal stressing rates from the
multi-cycle model can be scaled to provide ripe conditions for
earthquake rupture initiation in dynamic rupture models. We
compare our interseismic shear stressing rates to shear tractions
resolved from a uniform remote stress tensor used in the
dynamic rupture models. Shear stressing rates from the
interseismic model differ from the resolved shear tractions
indicating that fault interaction within the quasi-static impacts the
stress distribution. A comparison of the shear tractions estimated
using the interseismic shear stressing rates and time since last
ground-rupturing event show significant deviation from stresses
resolved from a remote stress tensor. This difference may greatly
impact dynamic rupture simulations. To further test the effect of
recent earthquakes, we simulate the 1992 Landers earthquake
to assess the impact of this rupture on the shear stress
distribution. Calculated absolute shear stress from the Landers
earthquake rupture differs from the absolute shear stress from a
model without the impact of the Landers earthquake by +- 0.4
MPa in right-lateral and reverse stresses and +- 0.6 MPa in
tensile stresses. Consequently, including nearby earthquake
ruptures varies the distribution of stresses and may provide a
more accurate stress distribution along the San Andreas fault.
We conclude that incorporating fault interaction, time since last
earthquake, and nearby earthquake ruptures in the stress
distribution along faults influences fault tractions that may
improve accuracy in dynamic rupture models earthquake
simulations.
Post-seismic ductile flow beneath the brittle-plastic
transition: constraints on rheology from microstructural
and electron backscatter diffraction (EBSD) analyses of
mylonitized pseudotachylite, South Mountains
metamorphic core complex, Arizona, Craig Stewart, and
Elena Miranda (Poster Presentation 078)
Post-seismic ductile flow promotes the upward rebound of the
brittle-plastic transition (BPT) after an earthquake, but it is difficult
to determine the rheologic controls on post-seismic flow due to
compositional and structural heterogeneities in the ductile middle
crust. We use microstructural and electron backscatter diffraction
(EBSD) analyses to evaluate the deformation mechanisms of
constituent minerals in pseudotachylyte-bearing granodiorite
mylonites from the footwall of the South Mountains core complex,
Arizona, to interpret the rheologic controls on post-seismic flow.
We focus on three pseudotachylyte-bearing samples developed
within an extensional shear zone associated with the core
complex detachment fault system, and they both entrain and are
variably overprinted by mylonitic fabrics, demonstrating multiple
episodes of seismic slip within ductilely flowing crust. All samples
show pseudotachylyte veins that are coplanar with or oriented at
low angles to mylonitic fabric, where the pseudotachylyte matrix
contains glass and ultrafine-grained mica, and clasts are
dominated by polycrystalline quartz and a paucity of feldspars.
The host granodiorite mylonite in all samples shows extensive
subgrain rotation recrystallization of quartz, and large,
microfaulted feldspar porphyroclasts with local bulging
recrystallization along fractures. The samples differ in that there
are variable amounts of polycrystalline quartz clasts in the
pseudotachylyte, and that some pseudotachylyte veins have
clasts with higher aspect ratios compared to other veins.
Recrystallized quartz microstructures in host mylonite are similar
to those in the polycrystalline clasts in pseudotachylyte,
indicating that pseudotachylyte preferentially formed in
recrystallized, quartz-rich foliation layers. However, some
polycrystalline quartz clasts exhibit 4-grain junctions, indicating
modification of microstructures during overprinting of
WELCOME
2016 SCEC Annual Meeting page 241
pseudotachylyte. These microstructures are consistent with a
quartz-controlled rheology of the crust below the BPT, where
dislocation creep in quartz is accompanied by the onset of grain
boundary sliding during post-seismic flow in mylonitized
pseudotachylyte. Our results suggest that post-seismic flow is
partitioned into the weaker, quartz-rich layers in the mylonites,
and that localized flow in weak layers could increase the rate at
which the BPT rebounds, promoting faster reloading of the brittle
fault and more frequent earthquakes.
Using Waveform Modeling to Determine the Depth of the
Deepest Earthquakes on and Near the Newport-
Inglewood Fault Zone in the Los Angeles Basin, Chloe S
Sutkowski, and Jascha Polet (Poster Presentation 227)
The depth extent of many faults in the Los Angeles Basin is still
poorly determined. It has been suggested that many are part of
shallow decollement systems. However, Boles et al. (2015)
conclude that a strong mantle helium signal along the Newport-
Inglewood Fault Zone indicates that it is a deep-seated fault,
directly or indirectly connected with the mantle. To help address
this question, we are analyzing seismic waveform data from
close-in broadband stations for the deepest earthquakes in the
LA basin above a magnitude 2.5 or higher in the last ten years.
Initial analysis of the relocated Southern California Seismic
Network catalog by Hauksson et al. (2012) reveals that for the
thirty-five deepest earthquakes in the LA basin, the majority
occurred near or along the Newport Inglewood Fault Zone at
depths in a range of 10 km - 17 km. The waveforms for close-in
stations for these events will be compared to synthetic
seismograms generated using a frequency-waveform approach,
for a suite of one-dimensional velocity models for the Los
Angeles Basin. Accurate depths will be determined by finding the
best match from these synthetics, calculated for a range of
depths, and the waveform data. We will present our preliminary
results from an analysis of several of these events. We will also
show maps and cross-sections of the earthquake focal
mechanism catalog by Yang et al. (2012) and the relocated
earthquake location catalog by Hauksson et al. (2012) for our
area of interest to investigate the correspondence to mapped
surface fault traces and the seismotectonic characteristics of the
Newport Inglewood Fault Zone along its length.
Fault creep observed on the Maacama and Rodgers
Creek faults, northern California using PS-InSAR, Jerlyn L
Swiatlowski, and Gareth J Funning (Poster Presentation 136)
Fault creep north of the San Francisco Bay Area has been
observed in a few discrete locations, along the Maacama and
Rodgers Creek faults, but the distribution of creep along these
faults are not mapped in detail. This is due to a high degree of
vegetation and minimal man-made structures surrounding the
faults. The Maacama fault is observed to creep at a four
alinement arrays along its entire 180 km extent (McFarland et al.,
2009). The Rodgers Creek fault has two more alinement arrays
than the Maacama fault as well as a previous InSAR study
showing fault creep around the city of Santa Rosa (Funning et
al., 2007).
Here we present persistent scatterer InSAR (PSI) results that
map creep on the Maacama and Rodgers Creek faults. We
processed a 39 image ERS descending track dataset, spanning
1992-2000, using the StaMPS PSI code (Hooper et al., 2004;
Hooper 2008), significantly increasing the density of surface
deformation observations along the southern Maacama fault
(track 113, frame 2817) and northern Rodgers Creek fault (track
113, frame 2835).
We identify fault creep along the Maacama fault surrounding the
cities of Ukiah and Willits, where creep has been detected at two
alinement arrays (McFarland et al., 2009). There is a change in
line-of-sight (LOS) velocity localized on the mapped fault trace,
consistent with right-lateral sense of fault motion. To compare the
LOS velocity on either side of the fault, the LOS velocity points
were projected onto profiles perpendicular to the fault, in
windows of 2 km long and 6 km wide. A best-fitting linear gradient
is subtracted and the change in velocity at the fault estimated by
fitting horizontal lines to data from either side of the fault. At Ukiah
and Willits, there are LOS velocity rates ranging from 0.1 – 1.6
mm/yr and 1.8 – 1.9 mm/yr, respectively. If projected into the fault
parallel direction, and assuming pure right-lateral strike-slip
motion, these rates are equivalent to creep rates of 0.3 – 4.1
mm/yr and 4.2 – 4.8 mm/yr, respectively.
Fault creep is identified on the Rodgers Creek fault surrounding
the city of Santa Rosa, CA, showing a change in LOS velocity
along the mapped fault trace where creep has been detected by
three alinement arrays (McFarland et al., 2009). This distribution
of creep is consistent with a previous study (Funning et al., 2007)
that identified right-lateral fault creep at rates of up to 6 mm/yr
between 1992-2001 using an ERS descending track dataset
(track 342, frame 2835). Additional LOS velocity estimates will be
created using other ERS tracks as well as using the small
baseline subset (SBAS) technique of StaMPS and processing
datasets from other satellites (e.g. Envisat, ALOS-1, ALOS-2,
Sentinel-1A and Sentinel-1B).
Interpreting Oligocene Paleogeography in Southern
California Using a Provenance Analysis of the Sespe
Formation, Carl Swindle, and Parker Clarke (Poster
Presentation 132)
The Sespe Formation is a lithified fluvial system that was
deposited during the Oligocene. Previous geological
investigations aimed at identifying the origins of the Sespe
Formation have arrived at conflicting conclusions. This project
assists in resolving these conflicts by identifying, mapping, and
writing lithologic descriptions of the Sespe that includes the
clastic variability and paleocurrent indicators. Data displayed
graphically from east to west, and cartographically using GIS
software is interpreted in order to suggest potential
paleogeography of the Oligocene. Analysis of data suggests the
locations of highlands present during the Oligocene, in particular,
two independent quartzose highlands along with other, subtler
findings. Understanding the dynamics of formations like the
Sespe is key to understanding the paleogeography of Southern
California.
Verification and Validation of High-Frequency (fmax = 5
Hz) Ground Motion Simulations of the 2014 M 5.1 La
Habra, California, earthquake, Ricardo Taborda, Kim B
Olsen, Robert W Graves, Fabio Silva, Naeem Khoshnevis,
William H Savran, Daniel Roten, Zheqiang Shi, Christine A
WELCOME
2016 SCEC Annual Meeting page 242
Goulet, Jacobo Bielak, Philip J Maechling, Yifeng Cui, and
Thomas H Jordan (Poster Presentation 272)
The Southern California Earthquake Center (SCEC) High-F
project seeks to advance physics-based, deterministic
earthquake simulation with the long-term objective of improving
ground motion prediction and seismic hazard models. An
important aspect of this involves the verification of models and
simulation methods, and the validation of synthetics with respect
to data. The present study describes a recent concerted effort to
verify different simulation methods, and attempts to reproduce
the ground motions from the 2014 M 5.1 La Habra, California,
earthquake in the greater Los Angeles region over a simulation
domain of 180 km x 135 km, and 62 km in depth. The simulations
are done using three high-performance computing codes; two of
them implement the finite difference method and the third one
implements a finite element approach. The models are tailored
to satisfy a maximum frequency of 5 Hz and a minimum shear
wave velocity of 500 m/s. We rely on the latest version of the
community velocity model CVM-S4.26.M01, and use a point
source and two finite source models. The point source model is
defined by a mechanism derived from strong-motion data and a
slip-time signal obtained from a dynamic rough-fault rupture
model. The two finite source models come from (i) an
independent source inversion study, and (ii) a simulation done
with a kinematic rupture generator also used in the SCEC
Broadband Platform. At the verification stage, we compared
synthetics from the three codes for the point source model, using
the 3D regional structure and frequency independent attenuation
(Q) models. The comparisons between the three codes exhibit
very good agreement. For the validation stage, we compared
seismograms collected at 300+ recording stations from different
regional strong-motion seismic monitoring networks with
synthetics obtained for the three different source models. Initial
results indicate that extended source models, even for a
moderate-size earthquake like the one considered here, tend to
lead to better fits with data. Current and future efforts concentrate
on defining the best possible source model based on goodness-
of-fit comparison metrics, and testing the influence of other
parameters such as a frequency dependent Q model and small
scale heterogeneities. Simulations have been carried out on
NCSA Blue Waters and on OLCF Titan.
Modeling the effect of roughness on the nucleation and
propagation of shear rupture, Yuval Tal, and Bradford H
Hager (Poster Presentation 051)
Faults are rough at all scales and can be described as self-affine
fractals. This deviation from planarity results in geometric
asperities and a locally heterogeneous stress field, which affect
the nucleation and propagation of shear rupture. We study this
effect numerically at the scale of small earthquakes, in which
realistic geometry and friction law parameters can be
incorporated in the model. We choose the minimum roughness
wavelength to be at the size of lab samples (10 – 20 cm) and
thus use the observed lab-scale slip or rate based friction laws
without modifying the constitutive parameters, assuming that the
experimental data already include the effects of smaller
wavelengths of roughness.
Our numerical approach includes three main features. First, to
enable slip that is large relative to the size of the elements near
the fault, as well as the variation of normal stress during slip, we
implement slip-weakening and rate-and state-friction laws into
the Mortar Finite Element Method, in which non-matching
meshes are allowed across the fault and the contacts are
continuously updated. In this method, Lagrange multipliers are
used to enforce the continuity of stress, the non-penetration
condition, and Coulomb friction on the fault in a weak integral
sense. Second, we refine the mesh near the fault using hanging
nodes, thereby enabling accurate representation of the fault
geometry. Finally, using a variable time step size, we gradually
increase the remote stress and let the rupture nucleate
spontaneously. This procedure involves a quasi-static backward
Euler scheme for the inter-seismic stages and a dynamic implicit
Newmark scheme for the co-seismic stages.
Simulations of 20-meter rough faults governed by velocity-
weakening rate-and state- friction show the significant effect of
roughness on: (1) Seismic moment; (2) Stress drop; (3) Slip rate;
and (4) Nucleation and propagation properties such as
nucleation length, rupture velocity, and breakdown zone.
Generally, under the same range of external loads, rougher faults
experience more events with smaller magnitudes and slip rates,
where the roughest faults experience only slow-slip aseismic
events. As the roughness amplitude increases, local barriers of
large normal stresses complicate the nucleation process and
result in asymmetric expansion of the rupture and larger
nucleation length. In the propagation phase of the seismic
events, the roughness results in larger breakdown zones.
Remotely triggered small earthquakes along the San
Jacinto Fault Zone, CA, Jennifer M Tarnowski, and Abhijit
Ghosh (Poster Presentation 239)
Although there are an increasing number of studies documenting
earthquakes triggered by dynamic stresses from teleseismic
events, there is still much to understand about earthquake
triggering on a regional scale. We focus on the central and
southern San Jacinto Fault Zone (SJFZ) region, which is one of
the most active fault zones in southern California. Previous
studies find little evidence for earthquakes near SJFZ triggered
by teleseismic events (Kane et al., 2007). However, it is not often
clear what method is best to evaluate triggering for this region.
We systematically examine the likelihood of remote triggering
along the SJFZ using multiple methods, including the β-value
statistic, temporal changes in peak magnitude, and visually
scanning seismic data. We systematically examine a ten-year
period with the β-value statistic, and find 25 candidate remote
teleseismic earthquakes that may have triggered small quakes
near SJFZ by dynamic stresses. After combining the results from
all of our methods, we conservatively estimate that 14 of the 25
candidates triggered small earthquakes near the SJFZ. We find
that it is difficult to use only the β-value statistic as an indicator of
triggering in the San Jacinto region. However, combining it with
the temporal change in peak magnitude can reveal earthquake
triggering with statistically significant results.
Temperature exerts the strongest control on the 3D
rheology of the southern California lithosphere, Wayne R
Thatcher, David S Chapman, and Colin Williams (Poster
Presentation 011)
Lithospheric temperature differences influence rheologic
behavior, crustal deformation and earthquake occurrence in
WELCOME
2016 SCEC Annual Meeting page 243
southern California. However, geotherms are poorly constrained,
with those used in published deformation modeling studies
typically differing by ~200˚-300˚C in the crust and upper mantle,
leading to factor of ~100 variations in estimates of effective
viscosity at these depths. Under SCEC auspices we are
developing a 3D Community Thermal Model (CTM) to improve
temperature estimates and so refine our understanding of
earthquake cycle and long-term deformation processes within
this complex but densely monitored and relatively well-
understood region.
We use more than 250 surface heat-flow measurements to
define 12 geographically distinct heat flow regions (HFRs). Model
geotherms within each HFR can be constrained by averages and
variances of surface heat flow q0 and thermal conductivity (k)
and radiogenic heat production (A). Using generic models for (A,
k) versus depth we have constructed preliminary 1D conductive
geotherms for each HFR. Regional temperature variations
between HFRs are as much as 300˚ in the mid-crust and 700˚ at
the Moho, much larger than the uncertainties in these preliminary
thermal calculations. Such temperature differences correspond
to several orders of magnitude lateral variations in effective
viscosity, surely large enough to have important implications for
modeling ductile deformation in southern California.
However, these local geotherm estimates and their uncertainties
can be improved with better constraints on rock type and
thermophysical properties within each HFR. The 1D (A, k)
structure can be more precisely estimated using inputs from (1)
surface geology and subsurface inferences on average
crust/upper mantle lithologic structure; (2) seismic imaging
results to determine average wave speed versus depth; (3)
empirical relations between seismic wave speed and rock type;
and (4) laboratory and modeling methods relating
thermophysical properties (A, k) to rock type.
A SCEC Community Velocity Model (CVM) that includes
independent estimates of S-wave velocity would permit better
correlation between seismic velocities and rock type.
Construction of a Geologic Framework Model for southern
California--explored in a June 2016 SCEC workshop--would
contribute importantly to more tightly constraining lithologic
structure within each HFR.
The Scale-Dependence of Fault Roughness and Asperity
Strength, Christopher A Thom, Emily E Brodsky, and David L
Goldsby (Poster Presentation 034)
The frictional properties of fault surfaces are controlled by the
collective behavior of asperity contacts. For a single asperity, the
force of friction is given by Ff = τa ∙ A, where τa is the shear
strength of an asperity and A is its contact area [Bowden & Tabor,
1950]. While a distribution of contact sizes is expected on a
frictional interface, the asperity shear strength is commonly
assumed to be constant. Recent work on the scale-dependent
roughness of natural faults, however, suggests that asperity
strength may also be scale-dependent [Brodsky et al., 2016]. In
order to directly test the link between scale-dependent fault
roughness and strength, we are conducting nanoindentation
tests over a wide range of indentation depths on a sample of the
Corona Heights Fault surface, whose roughness is well-
characterized down to nanometer length scales [Thom et al.,
2015]. For the Corona Heights Fault, roughness data collected
via atomic force microscopy suggest that indentation hardness,
which is proportional to yield stress, should decrease as a power-
law function of length scale as L-0.24. Nanoindentation tests
performed on non-geologic materials often show an ‘indentation
size effect’ (ISE), whereby hardness increases with decreasing
indentation depth [Pharr et al., 2010] in a manner consistent with
this prediction. To more generally investigate scale-dependent
strengths of geologic materials, we have performed
nanoindentation tests on single crystals of synthetic quartz and
San Carlos olivine to depths ranging from 30 to several hundred
nanometers. To date, these tests reveal a constant hardness with
no ISE for quartz. In contrast, the measured hardness of olivine
decreases with indentation size as a power law, with an exponent
in agreement with the value predicted above for a natural crustal
fault. If a similar scaling of material strength exists for faults like
Corona Heights, as revealed by nanoindentation of natural fault
surfaces, then our results will suggest that using a single value
of τa may not be appropriate when considering multi-asperity
friction on faults.
Constraints on the Source Parameters of Low-Frequency
Earthquakes in Parkfield and Cascadia, Amanda Thomas,
Gregory C Beroza, David R Shelly, Michael Bostock, Allan M
Rubin, Genevieve Savard, and Lindsey Chuang (Oral
Presentation 9/13/16 08:00)
Low-frequency earthquakes (LFEs) are small repeating
earthquakes that occur in conjunction with deep slow slip. Like
typical earthquakes LFEs are thought to represent shear slip on
crustal faults but when compared to earthquakes of the same
magnitude LFEs have lower corner frequencies, implying longer
durations, and are depleted in high-frequency content relative to
earthquakes of similar magnitude. Along the San Andreas Fault,
and in the Cascadia subduction zone, LFEs occur in rapid
succession, forming tectonic tremor. Here we present results
from LFE source studies in both Parkfield and Cascadia. In
Parkfield, we use an empirical Green’s function approach to
investigate which physical properties of the LFE source cause
high-frequency depletion. We find that the M~1 LFEs have typical
durations of ~0.2 s. Using the annual slip rate of the deep SAF
and the average number of LFEs per year we estimate average
LFE slip rates of ~0.24 mm/s. When combined with the LFE
magnitude this number implies a stress drop of ~104 Pa, two to
three orders of magnitude lower than ordinary earthquakes, and
a rupture velocity of 0.35 km/s. Typical earthquakes are thought
to have rupture velocities of ~80-90% the shear wave speed. In
Cascadia, we correct LFE waveforms for path effects and use
the resulting source time functions to calculate LFE duration and
magnitude. We use these estimates to show that, like Parkfield,
LFEs in Cascadia also have low stress drops, rupture and slip
velocities. We also find that LFE duration displays a weaker
dependence upon moment than expected for self-similarity,
suggesting that LFE asperities are limited in dimension and that
moment variation is dominated by slip. This behavior implies that
LFEs exhibit a scaling distinct from both large-scale slow
earthquakes and regular seismicity. Together the slow rupture
velocity, low stress drops, and slow slip velocity explain why
LFEs are depleted in high frequency content relative to ordinary
earthquakes and suggest that LFE asperities represent areas
capable of relatively higher slip speed in deep fault zones.
Additionally, changes in rheology may not be required to explain
WELCOME
2016 SCEC Annual Meeting page 244
both LFEs and slow slip; the same process that governs the slip
speed during slow earthquakes may also limit the rupture velocity
of LFEs.
Next Generation Boundary Element Models for
Earthquake Science, Thomas B Thompson, and Brendan J
Meade (Poster Presentation 338)
Modeling fault behavior and stress evolution in complex fault
geometries throughout the earthquake cycle is a fundamental
goal in earthquake science. To enable these studies we discuss
the development of Tectosaur, a boundary element library for
modeling crustal deformation that incorporates: 1) surface
topography, 2) gravity, 3) material property variations, and 4)
arbitrarily complex fault system geometry. A full-space Green's
function formulation using a novel kernel independent singular
integration method allows complete freedom in defining fault,
surface, and internal material interface geometries. A GPU-
centered implementation including approximation of far-field
element interactions using the fast multipole method admits
problems with hundreds of millions of elements. We discuss how
this boundary element approach can be used to implement
mechanically realistic block models and earthquake simulators.
Development and validation of an improved seismic
velocity model for the San Jacinto Fault region of
Southern California, Clifford H Thurber, Amir A Allam,
Xiangfang Zeng, Yan Luo, Hongjian Fang, and Haijiang Zhang
(Poster Presentation 213)
We are improving 3D body-wave and surface-wave tomographic
inversions in the San Jacinto Fault Zone (SJFZ) region, building
on previous research by (1) further expanding the arrival time,
phase velocity, and dispersion datasets, (2) extending the
tomography work to joint inversions of body-wave and surface-
wave data, and (3) assessing multiple 3D models in terms of high
frequency waveform data goodness-of-fit. We are expanding the
body-wave dataset by applying automatic arrival-time pickers to
more recent regional earthquake waveforms, and augmenting
the earthquake data with available seismic refraction data in the
region. The surface-wave dataset is being enlarged by extending
the range of dispersion data from ambient noise correlation
functions to both higher and lower frequencies. The original SJFZ
tomography studies inverted the body-wave and surface-wave
datasets separately, whereas we are carrying out joint inversions
of the two data types, yielding improved resolution and resulting
in increased waveform modeling accuracy. To quantify model
performance, we employ a time-frequency goodness-of-fit
method, decomposing 3-component signals into the time-
frequency domain and measuring the phase and amplitude
difference of the signal envelopes in each time-frequency
window. We compare synthetic waveforms to observed data from
moderate magnitude regional earthquakes simulated in several
tomographic models, including the SCEC CVM-H and previous
body-wave and surface wave models from separate inversions.
We find that all of the models reproduce body waves up to 3 Hz
with reasonable accuracy, but fail to match high-amplitude late-
arriving surface waves, especially in the 1-2 Hz range.
Nevertheless, our latest jointly-inverted model modestly but
consistently outperforms CVM-H in terms of modeled regional
waveform fits.
Accelerating AWP-ODC-OS Using Intel Xeon Phi
Processors, Josh Tobin, Alexander N Breuer, Charles Yount,
Alexander Heinecke, and Yifeng Cui (Poster Presentation 340)
AWP-ODC is software that simulates dynamic rupture and wave
propagation, using a staggered grid finite difference scheme, and
is widely used in the SCEC community. Recently a unified open
source version AWP-ODC-OS has been released. We present
an extension of AWP-ODC-OS to the Intel Xeon Phi processor,
codenamed Knight's Landing (KNL). Additionally, this extension
includes a new lightweight system that streamlines the execution
and verification of AWP-ODC-OS simulations.
The AWP-ODC code has in recent years run on GPU-
accelerated supercomputers. The hybrid version we have
developed allows AWP-ODC-OS to run additionally on Xeon Phi
processors, and is designed to take full advantage of the
potential of the KNL architecture. In particular, the KNL
processors provide 16GB of high-speed MCDRAM. Since the
time-to-solution of AWP-ODC is limited primarily by memory-
bandwidth, this high-bandwidth memory has the potential to
greatly reduce the time-to-solution of AWP-ODC. Moreover,
since even top-of-the-line GPUs are equipped with less than
16GB of memory, KNL presents the opportunity to run
simulations on larger domains than were previously possible. In
addition to the 16GB of MCDRAM, KNL nodes are outfitted with
up to 384GB of DDR memory, which presents further
opportunities to increase simulation size. Our extension
incorporates all of the features available in AWP-ODC-OS, which
include purely elastic kernels, frequency--dependent
viscoelasticity, free surface boundary conditions, absorbing
boundary conditions and point sources. We present our new
verification tool, which automatically generates a report with
synthetic seismograms and global velocity statistics after a
simulation, intended to allow the user to quickly verify the results
of a run.
Our performance results demonstrate that our extension of AWP-
ODC running on an Intel Xeon Phi 7210 processor performs
more than 7.5 times faster than an Intel Xeon E5-2680v3 and
more than 1.4 times faster than the recent Nvidia M40 GPU.
Postseismic deformation and viscosity of the Mexicali
region following the El-Mayor Cucapah earthquake
inferred from GPS observations, Xiaopeng Tong, Alejandro
Gonzalez-Ortega, Bridget R Smith-Konter, and David T Sandwell
(Poster Presentation 156)
It is important to study the deformation of the lithosphere to
understand the thickness, strength, and viscosity of the
lithosphere and substrate. Two northwest-southeast trending
fault systems connect the San Andreas Fault system to the rift
system in Baja California. The 2010 M7.2 El Mayor-Cucapah
earthquake ruptured these faults in the Mexicali region and
provides a unique observation to better understand the thickness
of the lithosphere and viscous substrate. We infer the viscosity
structure of the Mexicali region from an analysis of post-seismic
GPS time-series observations from 2010-2015, which consists of
campaign GPS sites surrounding the Sierra Cucaph and Indiviso
faults, and two survey lines covering the Cerro-Prieto fault and
the Imperial fault. We compare the post-seismic GPS time-series
data with a semi-analytic 3-dimensional viscoelastic model that
WELCOME
2016 SCEC Annual Meeting page 245
includes an elastic layer representing upper crust and a
viscoelastic substrate representing lower crust and upper mantle.
A coseismic slip model is used to generate stress changes
caused by the El Mayor-Cucapah earthquake. Here we focus on
the viscoelastic response of the postseismic transients to
estimate the thickness of the lithosphere and viscosity of the
substrate simultaneously. The horizontal component of GPS
data captured the large-scale postseismic transients that extend
tens to hundreds km from the fault, which provide critical
constraints to the viscoelastic model. Postseismic transients
observations from 2012-2015, two years after the earthquake,
suggests that the viscosity of the lower crust and upper mantle is
around 10^18 Pa s and the elastic layer is relatively thin in the
Mexicali region.
GrowClust: A hierarchical clustering algorithm for relative
earthquake relocation, with application to the Spanish
Springs and Sheldon, Nevada, earthquake sequences,
Daniel T Trugman, Peter M Shearer, and Ken D Smith (Poster
Presentation 200)
Accurate earthquake locations are essential for providing reliable
hazard assessments, understanding the physical mechanisms
driving extended earthquake sequences, and interpreting fault
structure. Techniques based on waveform cross-correlation can
significantly improve the precision of the relative locations of
event pairs observed at a set of common stations. Here we
describe GrowClust, a relative event relocation algorithm that can
provide reliable relocation results for earthquake sequences over
a wide range of spatial and temporal scales. The GrowClust
method uses input differential travel times, cross-correlation
values, and reference starting locations, and applies a hybrid,
hierarchical clustering algorithm to simultaneously group and
relocate events within similar event clusters. Requiring no explicit
matrix inversion, the method is computationally efficient and
capable of handling large data sets, and naturally applies greater
weight to more similar event pairs. Additionally, GrowClust
outputs robust location error estimates that are useful in
interpreting the reliability and resolution of relocation results. As
an example, we apply the GrowClust method to two prominent,
recent sequences in western Nevada: the Spanish Springs and
Sheldon earthquake swarms. These sequences highlight the
future potential for applying the GrowClust relocation method on
a much larger scale within Nevada and eastern California, where
existing relocation results are sparse but vital for understanding
the seismotectonics and seismic hazard of the greater Walker
Lane region.
Progress Report on Addition of a High-Speed Drive to
High-Pressure, Rotary-Shear Apparatus, Terry E Tullis
(Poster Presentation 030)
My rotary-shear, high-pressure machine at Brown University has
been operational for over 35 years. It allows unlimited slip on
annular samples at confining pressures up to 1 GPa. Nitrogen or
argon gas is used as the pressure medium in order to operate at
elevated temperatures and to use electronic transducers inside
the pressure vessel for measuring stress and displacement. The
samples are jacketed by a composite jacket assembly consisting
of O-rings to exclude the confining pressure from the sample and
Teflon rings to separate the O-rings from the sample so they will
not be torn apart by the large displacement and to transfer the
confining pressure to the sample. Separate pore pressure
systems access the top and bottom halves of the sample in order
to allow fluid flow through the sample, for measuring permeability
and for changing fluid composition during an experiment.
Rotation of the sample is driven by two systems. An
electrohydraulic stepping motor provides unlimited
displacements at over 7 orders of magnitude in speed from 0.001
microns/s to 10 mm/s, using two ~100:1 harmonic-drive speed-
reducers, one of which can be bypassed. Riding piggy-back on
the rotating table driven by this system is a rotary servo system
with 9 degrees of rotation. This is used for rapid rotary servo
control, allowing effective stiffening of the loading system, using
as the feedback device a resolver internal to the pressure vessel
mounted close to the sample’s sliding interface. The resolver has
24-bit resolution in one revolution, allowing a slip precision of 10
nanometers. The effective shear stiffness of the machine using
the resolver and rotary servo is increased 37x over the natural
stiffness to, for example, 1.8 MPa/micron at a normal stress of
25 MPa. This stiffness is only attainable at slip speeds below
about 3 microns/s, due to the finite response time of the
hydraulically actuated rotary servo.
I am in the process of adding a high-speed servo motor drive to
the apparatus to allow rotation of samples at slip speeds up to
~4.5 m/s. This will allow study of high-speed friction weakening
mechanisms at effective normal stresses similar to those at
which crustal earthquakes occur. The motor will be positioned on
the table rotated by the electrohydraulic stepping motor, taking
the place of the present hydraulic rotary servo drive. This will
allow continuous changes in speed of over 9.5 orders of
magnitude. The motor is an ETEL torque motor capable of
delivering over 100 kw of power. It can deliver 2250 Nm of torque
at speeds up to 400 rpm (1 m/s for our samples) and, using field
weakening, can rotate up to 1800 rpm, while delivering
progressively lower torques. The torque capability will allow
sliding samples with Byerlee friction values at normal stresses up
to 100 MPa. The motor’s torque and speed capabilities nicely
match the needs for our samples, meaning that no gearing is
required, eliminating that as a source of backlash. A 27-bit
Heidenhain absolute optical encoder will be mounted on the
motor as the feedback device for a Siemens motor controller,
allowing tight motor control to a precision corresponding to a
sample slip of 1 nanometer. The controller can use my existing
24-bit resolver in a secondary feedback loop, allowing the same
electronic stiffening of the apparatus as at present, but at speeds
up to 4.5 m/s.
Once the motor is installed and operational, jacketing of the
samples at high slip speeds will be the next challenge. Elevated
temperatures due to frictional heating of the samples will require
use of more refractory materials than the present Teflon rings
and Viton O-rings. Kalrez O-rings are good to 300 degrees C,
which, given their distance from the sample, may be sufficient for
the short duration of high-speed slip to realistic slips, e.g. 10 m
of slip at 2 m/s for 5 s. However, Teflon breaks down at 300
degrees C at room pressure and is closer to the sample so will
get hotter faster than the O-rings. Graphite, molybdenum
disulfide, and hexagonal boron nitride are refractory materials
that may be suitable to replace the Teflon, either alone or in some
composite geometry, perhaps involving Vespel, which is good to
500 degrees C. Only by fast sliding at high pressure, and some
WELCOME
2016 SCEC Annual Meeting page 246
trial and error, will it be possible to develop a satisfactory
refractory sliding jacket.
Toward the 3-component Crustal Motion Model:
Integration of Sentinel-1A SAR interferometry and
continuous GPS in the Los Angeles-Western Mojave
area, Ekaterina Tymofyeyeva, Homan Lau, and Yuri Fialko
(Poster Presentation 139)
The new Sentinel-1 mission, launched by the European Space
Agency in April 2014, provides extensive coverage at high spatial
resolution and frequent revisit intervals, which can dramatically
improve measurements of secular and transient deformation in
Southern California.
A full description of surface motion requires 3 orthogonal
components of the velocity field. InSAR data are limited to the
satellite line of sight, and cannot fully resolve 3-dimensional
deformation without simplifying assumptions. Continuous GPS
measurements are spatially more sparse than full-resolution
InSAR data, and suffer from high uncertainties in the vertical
direction, but provide robust estimates of horizontal velocities.
Together, InSAR and GPS measurements form a
complementary data set that can be used to resolve the 3-
dimensional deformation field. In this study, we combine data
from overlapping Sentinel-1 tracks 71 and 64 with two different
look directions, together with the horizontal components of
continuous GPS measurements, to estimate the 3-dimensional
secular velocity field in the Los Angeles basin and adjacent
Mojave desert.
We apply an iterative common point stacking approach to
estimate the contribution of the troposphere and ionosphere in
the InSAR data, and to isolate low-amplitude deformation signals
in our study region over the 2-year time period since the launch
of Sentinel-1A. To accommodate the differences in spatial
resolution between the InSAR and GPS data sets, we interpolate
the North and East continuous GPS measurements over the
areas in between the GPS sites. For each InSAR pixel, we use
the ascending and descending InSAR data, together with the
interpolated North and East GPS components, to solve for the
vertical and horizontal components of the velocity field.
High performance earthquake simulations in viscoelastic
media using SeisSol, Carsten Uphoff, Michael Bader,
Sebastian Rettenberger, and Alice-Agnes Gabriel (Poster
Presentation 341)
The software package SeisSol simulates seismic wave
propagation and dynamic rupture. Being based on unstructured
tetrahedral meshes, the software may be used in highly complex
scenarios with respect to fault geometry and heterogeneous
materials. The underlying discontinuous Galerkin method with
ADER time-stepping features high-order accuracy in space and
time and allows for an efficient implementation on modern
supercomputers. The software was shown to achieve petaflop
performance as well as high parallel efficiency on Xeon and Xeon
Phi based supercomputers in large scale dynamic rupture
scenarios. Furthermore, SeisSol's dynamic rupture
implementation has been verified using several SCEC test
cases, including rate-and-state friction laws, heterogeneous
initial conditions, and fault branching.
In order to achieve petascale, SeisSol's code base was heavily
changed to incorporate shared memory parallelization,
asynchronous communication, parallel I/O, and hardware aware
computational kernels. Until recently, the latter were only
available for elastic rheological models. We present our recent
efforts to extend the high performance kernels to support
viscoelastic attenuation. To this end, we have optimized and
extended our code generator for the innermost kernels. The
generator now automatically detects irrelevant matrix entries in
matrix chain products, determines zero-paddings for hardware
conformity, and enables block decompositions for structured
sparse matrices. We believe that these changes will simplify high
performance implementations of future extensions such as
anisotropy or poroelasticity.
The high performance implementation of attenuation was
validated on a plane wave problem, showing high-order
convergence, and a layer over halfspace problem (LOH.3). The
performance results indicate a similar efficiency as in the elastic
case. Furthermore, we simulated a dynamic rupture scenario
based on the 1994 Northridge earthquake using a power-law Q
model. Here, we achieve about 30% of peak performance on
14336 cores.
The Land-Atmosphere Interactions from Barometers and
Seismometers, Anne Valovcin, Jiong Wang, and Toshiro
Tanimoto (Poster Presentation 224)
Our understanding of seismic noise has improved dramatically in
the last 50 years. Also the cross-correlation approach now allows
us to make the Green’s functions and apply it to practical
applications. However, depending on the frequency band, the
cross-correlation approach does not necessarily work. Clearly,
this situation requires a better understanding of the nature of
seismic noise.
In this study, we focus on the direct interaction between the
atmosphere and the solid earth, or the land-atmosphere
interaction. The well-known microseisms (0.05-0.4 Hz) are
generated by ocean waves, but the land-atmosphere interactions
also generate an important part of low-frequency seismic noise,
such as tilt. There are now many stations that have both
barometers and seismometers from which we can study the
nature of the land-atmosphere interaction. We summarize our
findings from the local Southern California stations and the
Earthscope TA stations. For the latter stations, we examine two
aspects, one on the data when tropical cyclones hit the
Earthscope TA and the other is on the special subset of TA
stations that have had barometers, wind measurements and
other environmental sensors (temperature, humidity etc.).
We specifically focus on three aspects in our findings: (i)
Threshold pressure, (ii) tilt effects on horizontal components and
(iii) seasonal bifurcations that are seen at some stations.
The threshold pressure is easily recognized in vertical-
component data when they are plotted against pressure (from
co-located instruments). In Tropical Cyclone data, it is about
SP=10(Pa2/Hz) where SP is the power spectral density (PSD) of
surface pressure. Below this pressure value, vertical amplitudes
are basically constant and above this value seismic amplitudes
increase with pressure. This critical, threshold pressure can be
recognized at many stations.
WELCOME
2016 SCEC Annual Meeting page 247
In horizontal-component data, tilt dominates the signals, showing
increasing horizontal amplitudes with pressure for the entire
range of pressures. The effects of threshold pressure show up in
the changes of gradient in the plot of horizontal amplitudes vs.
pressure.
There are some stations for which horizontal amplitudes vs.
pressure show different behaviors from season to season; in
one-year long data set, we often see bifurcated patterns in the
amplitude vs. pressure plot; typically, in winter, horizontal
amplitudes show high constant values and do not change with
pressure. But in summer, horizontal amplitudes change with
pressure. This fact means that in summer atmospheric pressure
controls noise level, but in winter there is some noise generated
by another mechanism and because they are at higher
amplitudes, they dominate the horizontal-component data.
Induced earthquake magnitudes are as large as
(statistically) expected, Nicholas J van der Elst, Morgan T
Page, Debbie A Weiser, Thomas H Goebel, and Seyed M
Hosseini (Oral Presentation 9/13/16 14:00)
Injection-induced seismicity is a major contributor to seismic
hazard in the central US. The question now is whether induced
seismicity is amenable to statistical forecasting on any useful
timescale. One of the major uncertainties affecting such a
forecast is how large induced earthquakes can be. Are their
maximum magnitudes determined by injection parameters, or by
tectonics? Deterministic limits on induced earthquake
magnitudes have been proposed based on the size of the
reservoir or the volume of fluid injected. However, if induced
earthquakes occur on tectonic faults oriented favorably with
respect to the tectonic stress field, then they may be limited only
by the regional tectonics and connectivity of the fault network. In
this study, we show that the largest magnitudes observed at fluid
injection sites are consistent with the sampling statistics of the
Gutenberg-Richter distribution for tectonic earthquakes,
assuming no upper magnitude bound. The data pass three
specific tests: 1) the largest observed earthquake at each site
scales with the log of the total number of induced earthquakes,
2) the order of occurrence of the largest event is random within
the induced sequence, and 3) the injected volume controls the
total number of earthquakes, rather than the total seismic
moment. All three tests point to an injection control on
earthquake nucleation, but a tectonic control on earthquake
magnitude. The largest observed earthquakes are exactly as
large as expected from the sampling statistics. The results imply
1) induced earthquake numbers can be estimated from previous
activity and anticipated injection volumes, and 2) induced
earthquake magnitudes should be treated with the same
maximum magnitude bound that is used to treat seismic hazard
from tectonic earthquakes.
On nonstationarity corrections and durations in ground
motion applications of random vibration theory, Chris Van
Houtte, Tam Larkin, and Caroline Holden (Poster Presentation
275)
Random vibration theory (RVT) is a method for approximating
the peak time domain response of a signal from its Fourier
amplitude spectrum, based on assumptions of stationarity over a
portion of the signal duration, and random phase angles. In
ground motion prediction, RVT has primarily been used in the
stochastic method of ground motion simulation, to efficiently
generate 5%-damped response spectra from a model of the
earthquake Fourier amplitude spectrum. RVT methods are also
applied in geotechnical earthquake engineering, for 1D site
response analyses. This study examines the assumptions in
current applications of RVT, and investigates alternative ways of
improving the theoretical basis. Using recorded earthquake data
in the New Zealand Strong Motion Database, it is found that
correcting for an evolutionary power spectral density function
(PSDF) improves the method, however applying an evolutionary
bandwidth over-corrects the peak response prediction. The
definition of the signal duration is also reexamined. Rather than
using a 'Drms/Dgm' correction, as is common in the engineering
seismology literature, this study derives a model of the oscillator
duration directly. D5-75 is found to be the optimal duration metric
for prediction of peak response. It is found that, even with the
improved theoretical basis of the proposed method, there are still
long-period overpredictions of recorded response spectra using
RVT. The likely reasons behind these observations are also
presented in this study.
Status of ShakeAlert and G-FAST in the Pacific
Northwest, John E Vidale, Brendan W Crowell, J R Hartog, Paul
Bodin, Victor C Kress, and Ben Baker (Poster Presentation 237)
Here, we evaluate the real-time and historical performance of the
ElarmS‐2 algorithm in the Pacific Northwest. Real‐time and pre-
recorded seismic data from Oregon, California, and Washington
in the United States and British Columbia in Canada have been
used.
The prototype ShakeAlert system in the Pacific Northwest has
been operational since February 2013 with ElarmS-2. In
February 2015, a select group of local stakeholders began
receiving alerts. We have made several adjustments that make
the ElarmS‐2 algorithm less likely to falsely report multiple alerts
for earthquakes. Replaying of past earthquakes shows that the
new settings improve the algorithm’s behavior for edge cases
while not degrading previously well‐constrained solutions within
the dense part of the network. We expect few false alerts, and
none that would predict significant shaking anywhere in our
region. Even though ElarmS‐2 assumes a fixed earthquake
depth, the epicenter and magnitude estimates are accurate
enough to provide good predictions of shaking intensities even
for 50-km deep events.
In addition, the G-FAST GPS-based earthquake early warning
module has been in testing since September 2015. Starting in
early 2016, an Amazon Catalyst funded effort has been under
way to optimize and integrate G-FAST into ShakeAlert. We have
also created a test system that allows us to replay past and
synthetic earthquakes to identify problems with both the network
architecture and the algorithms, with options for altering latency,
data completeness, noise, and epicentral location. Events
replayed through the test system include the 2001 Mw 6.8
Nisqually, 2010 Mw 9.0 Tohoku-oki, 2010 Mw 8.8 Maule, 2014
Mw 8.2 Iquique and the 2015 Mw 8.3 Iquique earthquakes, and
we report the performance of G-FAST in rapidly modeling these
events.
WELCOME
2016 SCEC Annual Meeting page 248
Self-similar asymptotics of rate-strengthening faults,
Robert C Viesca, and Pierre Dublanchet (Poster Presentation
068)
We examine how slow slip progresses on rate-strengthening
faults. We consider that the source of rate-strengthening may be
a linear or non-linear viscous fault rheology, a logarithmic rate-
dependence, or a Dieterich-Ruina dependence on slip rate and
its history. We show the existence of self-similar asymptotic
solutions for slip rate of the form V = t^alpha f(x/t^beta). The
exponent beta is determined by the nature of the elastic
interaction (for slip between elastic half-spaces in contact, beta =
1; and for a layer sliding above a substrate, beta = 1/2). The
similarity exponent alpha is determined by the type of initial or
boundary conditions. Such conditions may be, for example, (i) a
sudden change in stress on the fault or (ii) an imposed boundary
slip rate. We consider in-plane or anti-plane slip for examples (i)
and (ii) and present the asymptotic solutions thereof, which may
be found numerically or in closed form. The self-similar behavior
of scenario (i) is, for a step increase in stress, that of an initially
elevated slip rate decaying in time while spreading in space; and
of scenario (ii) is that an elevated slip rate propagating along the
fault. Under scenario (i) we show that the disparate fault
rheologies share a common closed-form similarity solution for the
decay of slip rate following the initial stress change. For
comparison, we compute numerical solutions to the evolution
equation for slip rate (and state, when applicable) and find
precise agreement with the above analysis.
Using Coupled Geomechanical Modeling to Investigate
Potential Production Induced Seismicity, Robert L Walker,
Susan E Hough, Birendra Jha, Victor C Tsai, Morgan T Page,
and Fred Aminzadeh (Poster Presentation 345)
Recent work has suggested that oil and gas industry operations
may have triggered significant induced seismicity in several
cases throughout the twentieth century. We have chosen one
historic example to apply a coupled fluid flow and geomechanical
framework in order to investigate the potential that oil production
may have resulted in significant effect.
We consider the Mw7.3 1952 Kern County earthquake, and the
suggestion that it may have been induced by production in the
Wheeler Ridge oil field. The commencement of production from
deep Eocene production horizons within the Wheeler Ridge field
preceded the mainshock by 111 days, with the producing
formation located at depths reaching three kilometers, within one
kilometer of the White Wolf fault. While the epicentral location of
the earthquake is not precisely known, it is thought to be within
seven kilometers of the wellhead of the initial producing well from
the deep formation.
Given this circumstantial evidence, as well as parameters guided
by available industry records and a knowledge of traditional oil
industry practices, we extend our investigation by means of
petroleum industry standard numerical reservoir simulation, and
by means of a numerical simulation coupling both fluid flow and
poroelastic effects. Along with the study of this specific case, we
seek to use this framework to better understand and identify
potential production related triggering mechanisms, as well as
examine industry both past and present that could have an effect
on induced seismic hazard.
The California Plate Boundary Observatory GPS-GNSS
Network, Christian Walls, Doerte Mann, Ryan C Turner, Andre
Basset, Shawn Lawrence, Kenneth Austin, Stephen T Dittman,
Karl Feaux, and Glen S Mattioli (Poster Presentation 144)
The EarthScope PBO GPS-GNSS network in California, funded
by the NSF and operated by UNAVCO, is comprised of 599
permanent GPS and GNSS stations spanning three principal
tectonic regimes and is administered by separate management
regions (Subduction - Pacific Northwest [91 sites], Extension -
East [41 sites], Transform - Southwest [467 sites]). Since the
close of construction in September 2008 various enhancements
have been implemented through additional funding by the NSF,
NOAA, and NASA and in collaboration with stakeholders such as
Caltrans, Scripps, and the USGS.
Initially, the majority of stations used first generation IP based
cellular modems and radios capable of ~10KB/s data transfer
rates. Aside from 1-2MB daily downloaded files the bandwidth
limitation was a challenge for regional high-rate data downloads
for GPS-seismology and airborne LiDAR surveys, and real-time
data flow. Today, only 5 of the original cell modems remain with
338 upgraded cell modems providing 3G/4G/LTE data
communications with transfer rates ranging from 80-400 KB/s.
Ongoing radio network expansion and upgrades continue to
harden communications using the 900MHz, 2.4GHz and 5.8Ghz
spectrums. 27 VSAT and one manual download site remain
(down from 60 VSAT & 12 Manual in 2008). In California, the
network capabilities for 1Hz & 5Hz downloads or low latency 1
Hz streaming are ~95%, ~85% and ~70%, respectively. In
general, nearly all custom data requests for high rate data are
accommodated without delay from communications limitations.
During the past year, uptime ranged from 95-100% with data
return for 15 s data exceeding 99%. 200 additional California
sites were added in March 2016 to the real-time (1Hz) system
and data from 410 sites are distributed in BINEX and RTCM
2.3/3.1 formats with an average latency of ~0.7 s and completion
of ~85%. These include some challenging locations such as the
Channel Islands and the Sierras. A variety of geophysical
sensors are co-located with a subset of the stations and include:
21 MEMS accelerometers, 31 strong motion and broadband
seismometers, 9 borehole strainmeters and 1 long baseline
strainmeter. Vaisala meteorological instruments are located at
53 sites of which 52 stream GPS/Met data. As budget allows,
GPS-only receivers are replaced with GNSS receivers and
antennas. Today, 120 stations in California are GLONASS
enabled NetR9 receivers or full constellation Septentrio PolaRx5
receivers.
Observations and Models of Co- and Post-Seismic
Deformation Due to the 2015 Mw 7.8 Gorkha (Nepal)
Earthquake, Kang Wang, and Yuri Fialko (Poster Presentation
158)
The 2015 Mw 7.8 Gorkha (Nepal) earthquake occurred along the
central Himalayan arc, a convergent boundary between India
and Eurasian plates. We use space geodetic data to investigate
co- and post-seismic deformation due to the Gorkha earthquake.
Because the epicentral area of the earthquake is characterized
by strong variations in surface relief and material properties, we
developed finite element models that explicitly account for
WELCOME
2016 SCEC Annual Meeting page 249
topography and 3-D elastic structure. Compared with slip models
obtained using homogenous elastic half-space models, the
model including elastic heterogeneity and topography exhibits
greater (up to ~ 10%) slip amplitude. GPS observations spanning
more than 1 year following the earthquake show overall
southward movement and uplift after the Gorkha earthquake,
qualitatively similar to the coseismic deformation pattern.
Kinematic inversions of GPS data, and forward modeling of
stress-driven creep indicate that the observed post-seismic
transient is consistent with afterslip on a down-dip extention of
the seismic rupture. The Main Himalayan Thrust (MHT) has
negligible creep updip of the 2015 rupture, reiterating a future
seismic hazard. A poro-elastic rebound may contribute to the
observed uplift southward motion, but the predicted surface
displacements are small (on the order of 1 cm or less). We also
tested a wide range of visco-elastic relaxation models, including
1-D and 3-D variations in the viscosity structure. All tested visco-
elastic models predict the opposite signs of horizontal and
vertical displacements compared to those observed. Available
surface deformation data allow one to rule out a model of a low
viscosity channel beneath Tibetan Plateau invoked to explain
variations in surface relief at the plateau margins.
Using coda waves to resolve the scattering and intrinsic
attenuation structure of Southern California, Wei Wang,
and Peter M Shearer (Poster Presentation 205)
Measuring intrinsic and scattering attenuation is important for a
variety of geophysical applications. Charactering scattering and
absorbing properties and the power spectrum of crustal
heterogeneity is a fundamental problem for informing strong
ground motion estimates at high frequencies, where scattering
and attenuation effects are critical. We perform a comprehensive
study of coda behavior in Southern California to constrain
scattering and intrinsic attenuation structure. We stack 31,894
vertical-component envelopes and 29,203 transverse-
component envelopes at 2 to 4 Hz from 1224 spatially distributed
earthquakes from 1981–2013 at source depths of 10 to 15 km
and epicentral distances from 0–250 km with magnitudes larger
than 1.8. The stacked envelopes are averaged as a function of
distance, depth and frequency. We model these observations
using a particle-based Monte Carlo algorithm to simulate regional
scattering and intrinsic attenuation. This method has the
advantage of including both P- and S-wave scattering and both
single and multiple scattering events. The synthetic results show
that coda waves for the Southern California region can be
reasonably fit with a two-layered model, composed of a shallow
crustal layer with strong wide-angle scattering and high intrinsic
attenuation and a deeper layer with weaker scattering and lower
intrinsic attenuation (top 5.5 km: QP = 250, QS = 125,
heterogeneity correlation length a = 50 m, velocity heterogeneity
rms e = 0.5; lower crust: QP = 900, QS = 400, a = 2 km, e = 0.05).
San Diego Earthquake Hazard: Geotechnical Data
Synthesis, Luke Weidman, Jillian M Maloney, and Thomas K
Rockwell (Poster Presentation 093)
With a population of ~1.3 million, the City of San Diego is the third
largest city in California, and it is traversed by the Holocene-
active Rose Canyon Fault Zone (RCFZ). The Rose Canyon Fault
is a strike-slip fault with a slip rate of 1-2 mm/yr and the potential
to produce a M6.9 event. This project focuses on the strands of
the RCFZ that traverse through the downtown area, which is the
economic center of the city. The seismic hazard of the RCFZ has
a direct impact on development in and around the city via the
Alquist-Priolo Earthquake Fault Zoning Act, which regulates
locations of structures for human occupancy. As a result,
geotechnical firms in San Diego have been conducting many
private, small-scale studies and investigations of the local fault
architecture since the 1980s and have amassed an impressive
amount of data. However, each report is plot or parcel specific
and, at most, will only reference data from neighboring parcels.
As there exists no resource where all of the data can be studied
at once, reports are commonly studied independently. This
project synthesizes the existing geotechnical data into a
comprehensive geodatabase in an effort to show the current fault
geometry within the city and provide insight to the evolution of
the RCFZ.
Historically, geotechnical companies are hesitant to share results
of fault studies and field investigations with each other, as the
data is proprietary. Recently, however, several geologists and
engineers within the professional community have called for a
combined resource and many of San Diego’s geotechnical firms
have contributed data, including Geocon, Kleinfelder, Leighton &
Associates, CTE, and URS/AECOM. This project compiles the
data as a means to contribute to and improve upon the
community fault model and give the city an accurate model for
use in updating its seismic safety element. The geodatabase also
aids the science community by helping to establish the variety of
fault characteristics and complexities along strike, illuminate
recurrence intervals or patterns of multi-segment ruptures, and
provide evidence for long term slip rate. To date, we have
collected over 500 geotechnical reports, and locations of all
trenches, borings, and CPT soundings are being compiled in a
GIS database.
Science-based decision making in a high-risk energy
production environment, Debbie Weiser (Poster Presentation
180)
Energy production practices that may induce earthquakes
require decisions about acceptable risk before projects begin.
How much ground shaking, structural damage, infrastructure
damage, or delay of geothermal power and other operations is
tolerable? I review a few mitigation strategies as well as existing
protocol in several U.S. states. Timely and accurate scientific
information can assist in determining the costs and benefits of
altering production parameters. These issues can also be
addressed with probability estimates of adverse effects (“costs”),
frequency of earthquakes of different sizes, and associated
impacts of different magnitude earthquakes.
When risk management decisions based on robust science are
well-communicated to stakeholders, mitigation efforts benefit.
Effective communications elements include a) the risks and
benefits of different actions (e.g. using a traffic light protocol); b)
the factors to consider when determining acceptable risk; and c)
the probability of different magnitude events.
I present a case example for The Geysers geothermal field in
California, to discuss locally “acceptable” and “unacceptable”
earthquakes and share nearby communities’ responses to
smaller and larger magnitude earthquakes. I use the USGS’s
“Did You Feel It?” data archive to sample how often felt events
WELCOME
2016 SCEC Annual Meeting page 250
occur, and how many of those are above acceptable magnitudes
(to both local residents and operators). Using this information, I
develop a science-based decision-making framework, in the
case of potentially risky earthquakes, for lessening seismic risk
and other negative consequences. This includes assessing
future earthquake probabilities based on past earthquake
records. One of my goals is to help characterize uncertainties in
a way that they can be managed; to this end, I present simple
and accessible approaches that can be used in the decision
making process.
CSEP Evaluations of 24-Hour Earthquake Forecasting
Models for California: New Results and Ensemble
Models, Maximilian J Werner, Maria Liukis, Warner Marzocchi,
David A Rhoades, Matteo Taroni, Zechar D Zechar, and Thomas
H Jordan (Poster Presentation 314)
Operational Earthquake Forecasting requires credible
earthquake probability estimates at short time scales. The
objective of the Collaboratory for the Study of Earthquake
Predictability (CSEP) is to evaluate earthquake forecasting
models and hypotheses in a blind, automated and prospective
manner. CSEP supports OEF efforts by independently and
rigorously evaluating the strengths and weaknesses of candidate
OEF models and ensemble OEF models. CSEP has been
evaluating over a dozen 24-hour forecasting models in California
since 2009. Models include the STEP model, various ETAS
model flavors, non-parametric models and other statistical
clustered seismicity models. Here, we report on new results from
CSEP’s 24-hour earthquake forecasting experiment in California.
The data set consists of 132 earthquakes greater than magnitude
3.95. Relative probability gains indicate that the predictive skills
of the recent models are improving. In addition, we explore
protocols for constructing ensemble models. Ensembles are
powerful forecasting tools when several models are available and
their predictive skills not yet established. In addition,
combinations of models may outperform individual models.
Geological Observations on History and Future of Large
Earthquakes along the Himalayan Frontal Fault Relative
to the April 25, 2015 M7.8 Gorkha Earthquake near
Kathmandu, Nepal, Steven G Wesnousky (Poster
Presentation 095)
Steven G. Wesnousky1, Yasuhiro Kumahara2, Deepak
Chamlagain3, Ian Pierce1, Alina Karki3, Dipendra Gautam4: 1
Center for Neotectonic Studies and Seismological Laboaratory,
University of Nevada, Reno 89557, USA, 2 Graduate School of
Education, Hiroshima University, 1-1-1, Kagamiama, Higashi-
Hiroshima, Hiroshima 739-8524, Japan, 3 Department of
Geology, Tri-Chandra Multiple Campus, Kathmandu, Nepal, 4
Centre for Disaster and Climate Change Studies, Kathmandu,
Nepal
The 2015 Gorkha earthquake earthquake produced
displacement on the lower half of a shallow decollement that
extends 100 km south, and upward from beneath the High
Himalaya and Kathmandu to where it breaks the surface to form
the trace of the Himalayan Frontal Thrust (HFT), leaving
unruptured the shallowest ~50 km of the decollement. To
address the potential of future earthquakes along this section of
the HFT, we examine structural, stratigraphic, and radiocarbon
relationships in exposures produced by emplacement of
trenches across the HFT where it has produced scarps in young
alluvium at the mouths of major rivers at Tribeni and Bagmati.
The Bagmati site is located south of Kathmandu and directly up
dip from the Gorkha rupture, whereas the Tribeni site is located
~200 km to the west and outside the up dip projection of the
Gorkha earthquake rupture plane. The most recent rupture at
Tribeni occurred 1221 – 1262 AD to produce a scarp of ~7 m
vertical separation. Vertical separation across the scarp at
Bagmati registers ~10 m, possibly greater, and formed between
1031 - 1321 AD. The temporal constraints and large
displacements allow the interpretation that the two sites
separated by ~200 km each ruptured simultaneously, possibly
during 1255 AD, the year of a historically reported earthquake
that produced damage in Kathmandu. In light of geodetic data
that show the equivalent of ~20 mm of crustal shortening across
the HFT occurs on an annual basis, the sum of observations is
interpreted to suggest that the HFT extending from Tribeni to
Bagmati may rupture simultaneously, that the next great
earthquake near Kathmandu may rupture an area significantly
greater than the section of HFT up dip from the Gorkha
earthquake, and that it is prudent to consider that the HFT near
Kathmandu is approaching or in the later stages of a strain
accumulation cycle prior to a great thrust earthquake, most likely
much greater than occurred in 2015.
Aftershock productivity of large megathrust earthquakes:
regional variations and influence of mainshock source
parameters, Nadav Wetzler, Emily E Brodsky, and Thorne Lay
(Poster Presentation 169)
Aftershock productivity is observed to increase with mainshock
magnitude following a well-defined relationship for any given
region. However, variation of this scaling relationship is still
poorly characterized. We focus on variations of aftershock
productivity of large (MW ≥ 7.0) circum-Pacific megathrust
earthquakes within the past 25 years. We investigate how
aftershock productivity varies regionally and with respect to
mainshock source parameters determined by finite-fault rupture
model inversions. Aftershock productivity is higher for islands
arcs of the western circum-Pacific than for continental arcs of the
eastern circum-Pacific. Events with larger static stress-drop tend
to produce fewer aftershocks than comparable magnitude events
with lower stress-drop. The regional trend may reflect differences
in fault susceptibility between island arcs and continental arcs,
whereas the stress-drop behavior may imply near-total stress-
drop release for some events.
Comparisons between the 2016 USGS induced-
seismicity hazard model and “Did You Feel It?” data,
Isabel White, Taojun Liu, Nicolas Luco, and Abbie B Liel (Poster
Presentation 247)
The steep increase in seismicity rates in Oklahoma and southern
Kansas led the U.S. Geological Survey (USGS) to develop a one-
year induced seismicity probabilistic seismic hazard model for
2016 (Petersen et al., 2016). The increase has also led to a
relatively large amount of ground motion and macroseismic
intensity observations. In order to ground truth the model results,
two databases were compared with the model: peak ground
acceleration (PGA) values from instrumental data and intensity
data from the “Did you feel it”? (DYFI?) system. As the 2016
WELCOME
2016 SCEC Annual Meeting page 251
hazard model was heavily weighted to the catalogs from 2014-
2015, this study focuses on data from these years. The process
involved converting hazard model results and any available
ground motion data from PGA to Modified Mercalli Intensity
(MMI) for comparison with the DYFI? data. A number of available
conversions were tested; the one by Atkinson and Kaka (2007),
which includes Central and Eastern U.S. data, proved to be best.
The seismic data had to be declustered and the hazard model
adjusted for site conditions to be more comparable. The annual
exceedance rates were calculated with the DYFI? data for
comparison to the model. The comparison was complicated by a
number of issues related to the spatial and temporal
completeness of the DYFI? data as described by Mak and
Schorlemmer (2016). Furthermore, whereas the hazard model
focuses on a higher range of ground motion, usually above about
MMI IV or so, to date there are limited DYFI? data above IV from
Oklahoma earthquakes. Most of the DYFI? responses are in the
MMI II-IV range. Despite these limitations, comparisons at a
number of key sites in the area show agreement in the annual
rates of exceedance around MMI IV.
A detailed automatic 1998-2015 earthquake catalog of the
San Jacinto fault zone region, Malcolm C White, Zachary E
Ross, Frank L Vernon, and Yehuda Ben-Zion (Poster
Presentation 174)
The Anza (AZ) network has collected continuous seismic data
since 1998 focusing on the San Jacinto Fault Zone (SJFZ).
Nearby stations of the regional CI network and several
deployments (PB, SB and YN) around the SJFZ provide
important additional data. We discuss a seismic catalog derived
from a virtual SJFZ network combining stations of the above
networks. An accurate and consistent catalog extending to low
magnitudes is fundamental for many investigations of
earthquakes and structures. While analyst reviewed hypocenters
produce the most accurate catalogs, the required analyst hours
for the large dataset of the virtual SJFZ network is prohibitively
high. In addition, analyst reviewed catalogs produced over
decades often contain inconsistencies as various analysts
inevitably have differing decision making processes and may
apply different tools as they become available. An automated
procedure can analyze very large data sets and offers a uniformly
consistent product. Special care, however, must be taken to
ensure that the accuracy of results are reliable and that the level
of completeness of the catalog is understood. Any systematic
biases must also be understood to effectively use the catalog for
further study. Here we develop a detailed earthquake catalog for
the region using an automatic workflow that includes the
following key processing steps: (i) A newly developed S-wave
detection algorithm, (ii) location inversion using a 3D velocity
model derived from joint inversion of surface and body waves,
and (iii) double-difference relocation algorithm. Hypocenter
errors are estimated using a bootstrap resampling method and
moment magnitudes are estimated from the low-frequency
asymptote of stacked displacement spectra obtained via multi-
taper spectral analysis. Example aspects of the catalog will be
presented in the meeting.
Testing the shorter and variable recurrence interval
hypothesis along the Cholame segment of the San
Andreas Fault, Alana M Williams, Ramon Arrowsmith, Sinan O
Akciz, Thomas K Rockwell, Lisa Grant Ludwig, and Sally
Branscomb (Poster Presentation 128)
The Cholame segment of the San Andreas Fault interacts with
the Parkfield segment to the northwest with its creep and M6
earthquakes, and the locked Carrizo segment to the southeast.
Although offset reconstructions exist for this ~75 km reach,
rupture behavior is poorly characterized, limiting seismic hazard
evaluation. Here we present new paleoseismic results from 2
fault perpendicular 26 m long trenches connected by a 15 m long
fault parallel trench. The site is located south of the Parkfield
segment 20 km southeast of Highway 46. Site geomorphology is
characterized by several ~50 m offset drainages northwest of the
trenches, small shutter ridges, sag ponds, and alluvial fans
crossing the fault. Fault zone stratigraphy consists of alternating
finely bedded sands, silts, and gravels, and bioturbated soil
horizons. The strata record 3-4 earthquakes and possible
deformation post-1857, similar to the LY4 site 38 km southeast.
E4, E3 and the most recent earthquake (MRE) are well supported
by evidence of decreasing vertical offset up-sequence, capped
fissure fill and colluvial wedges, which create small horst and
graben structures. Units display vertical offsets ranging from 60
cm at the base to 12 cm near the MRE horizon, small colluvial
wedges, and sag deposits within the ~4 m wide fault zone. E2—
the penultimate--is less certain, supported only by the decreasing
offset in the stratigraphic sequence. The E4 event horizon is a
gradational clayey silt sag deposit capped by discontinuous
gravel, 18 cm at its thickest point and extending 4.8 m across the
fault zone. The E3 and E2 event horizons are capped by thin
bedded silty clay, and bounded by discontinuous burn horizons.
The MRE horizon extends 6 m across the main fault zone, and
consists of a silty clay sag deposit capped by very fine, bedded
sand and coarse gravel, 22 cm at its thickest point and overlying
a burn horizon. If the MRE is indeed the 1857 event, it has
significant potential in correlation with the high quality rupture
records at Bidart (70 km southeast), and Frazier Mountain (180
km southeast). This site contains abundant detrital charcoal in
many of the units and burn horizons at or near event horizons
providing great potential for bracketing the age of these
paleoearthquakes.
Using Virtual Quake for GNSS tsunami early warning:
coupling earthquakes, tsunamis, and ionosphere physics,
John M Wilson, Kasey W Schultz, John B Rundle, and Ramya
Bhaskar (Poster Presentation 318)
In light of growing demand for fast Global Navigational Satellite
System tsunami early warning in the Pacific Rim, computer
simulations are being developed to play key roles at all stages.
Virtual Quake is an earthquake simulator that can rapidly
produce long histories of earthquakes for complex fault systems,
including tsunamigenic regions, as well as expected seafloor
uplift from such earthquakes. When paired with the Tsunami
Squares simulator, large catalogs of tsunami scenarios can be
generated. To translate potential tsunamis to potential
observables, we plan on coupling our tsunami generation
pipeline to a simulation of total electron content (TEC)
fluctuations in the ionosphere caused by tsunamis. These TEC
signatures can be detected by GPS receivers, allowing fast
comparison of ionospheric perturbations to precomputed
catalogs of tsunami scenarios for rapid hazard assessment.
WELCOME
2016 SCEC Annual Meeting page 252
Validation of Deterministic Broadband Ground Motion and
Variability with Ground Motion, Kyle B Withers (Poster
Presentation 178)
The decay of energy at high frequencies, decay as a function of
distance, source model, and complementary relations between
scattering and apparent attenuation are all important features
that simulations need to accurately capture to be used for
engineering purposes. With the recent addition of realistic fault
topography in 3D simulations of earthquake source models,
ground motion can be deterministically calculated more
realistically up to higher frequencies. Here, we model dynamic
rupture propagation for both a generic strike-slip event and blind
thrust scenario earthquakes matching the fault geometry of the
1994 Mw 6.7 Northridge earthquake along rough faults up to 7.5
Hz. We include frequency-dependent attenuation via a power law
above a reference frequency in the form Q0fn and include
nonlinear effects via Drucker-Prager plasticity. We model the
region surrounding the fault with and without small-scale medium
complexity in both a 1D layered model characteristic of southern
California rock and a 3D medium extracted from the SCEC CVM-
4.26 model including a near-surface geotechnical layer.
However, before the broadband deterministic simulations can be
used for engineering applications, validation is required to
demonstrate that the energy decay in the synthetics as a function
of distance is similar to that for observed ground motions. Here,
in addition to comparing with GMPEs, we compare with simple
proxy metrics to evaluate the performance of our deterministic
models and to determine the importance of the different
complexities within our model. We find that 3D media
heterogeneity, at both the long and short scale, is necessary to
agree with data, and should be included in future simulations to
best model the earthquake ground motion and variability. We
also measure the decay of energy at high frequencies, known as
κ, and find realistic results across a narrow bandwidth. We
introduce a shallow layer with frequency-independent Q layer in
the near surface that serves to simulate the site-effect of κ; this
technique may become necessary as simulations continue to
extend to higher frequencies.
Characterizing emissivity spectra from geomorphic
surfaces along the southern San Andreas Fault, Ryan D
Witkosky, Paul M Adams, Kerry Buckland, Kenneth W Hudnut,
Patrick D Johnson, David K Lynch, Katherine M Scharer, Joann
M Stock, and David M Tratt (Poster Presentation 085)
Geologic mapping and cosmogenic exposure ages of alluvial
surfaces along the southern San Andreas Fault provide
independent data sets to evaluate corresponding emissivity
spectra from thermal hyperspectral airborne imagery. We use
new 1-m pixel resolution data from 2015 covering the Mission
Creek strand between Thousand Palms Oasis and Pushawalla
Canyon, where a set of alluvial surfaces are progressively offset
by the fault. After converting the image data from at-sensor
radiance to emissivity, we examined spectra from each of four
dated surfaces that span ~10 to 90 ka (Blisniuk and Sharp, 2014).
This comparison reveals that the magnitude of an absorption
feature located between 9-10 micron wavelengths generally
increases with age of the surface. We hypothesize that the depth
of this absorption feature varies by exposure age as clay
minerals accumulate in coatings of desert varnish on the surface.
Lab spectra for clays commonly found in desert varnish (e.g. illite
and montmorillonite) show a significant absorption feature
between 9-10 micron wavelengths caused by vibrational
stretching in the silicon-oxygen tetrahedra that comprise the
crystal lattice. In combination, these results suggest that as these
geomorphic surfaces alter with age, accretion of desert varnish
results in a greater contribution to the spectrum for each pixel.
Using varnish development as an age proxy for relative dating is
common in the literature but results have generally been
inconclusive or difficult to apply more broadly due to regional and
temporal changes in surface development. We plan to further
analyze the emissivity spectra at specific sites to determine
whether the observed spectral absorption feature behaves
consistently or if sampling strategies alter the results. If the
preliminary observations are robust, this technique may provide
a tool for expanding existing age data, or reconnaissance
evaluation of surface ages in arid locales that are conducive to
varnish development, facilitating slip rate studies and guiding
dating efforts for these surfaces.
Off-fault plasticity in dynamic rupture simulations: 3D
numerical analysis and effects on rupture transfer in
complex fault geometries, Stephanie Wollherr, and Alice-
Agnes Gabriel (Poster Presentation 062)
Estimating the potential dimension of a future earthquake hosted
by a complex fault system depends crucially on rupture
branching and jumping dynamics between adjacent fault
segments. Numerical modelling of such earthquake source
dynamics requires realistic physical assumptions, such as the
incorporation of plastic yielding around the fault induced by high
stresses at the rupture tip. Further, high-resolution numerical
methods are needed for a precisely capturing of on- and off-fault
dynamics in conjunction with seismic wave propagation.
Here we present three-dimensional dynamic rupture simulations
with off-fault plasticity focusing on numerical convergence and
the influence on rupture transfers between fault segments in an
application scenario. The simulations are performed with SeisSol
(www.seissol.org), an open-source software package based on
the high-order ADER-DG method solving frictional sliding
coupled to seismic wave propagation. We first verify the (visco-)
plastic implementation by benchmark tests of the SCEC
Spontaneous Rupture Code-Validation project. To demonstrate
the consistency of this nonlinear rheology in a 3D modal DG
approach we perform in-depth on-fault convergence tests. Our
results show that plasticity regularizes a mesh-dependency of
peak slip rate observed in purely elastic simulations. On the other
hand, plasticity reduces the required mesh discretization for a
consistent rupture arrival time compared to elastic simulations.
The ADER-DG method is especially suited for complex
geometries by allowing the usage of unstructured tetrahedral
meshes. In this context, we investigate how plastic yielding
influences the rupture propagation in a large-scale dynamic
rupture scenario based on the 1992 Landers earthquake. Our
numerical model reproduces well the total rupture duration and
the seismic moment of the event. We observe complex rupture
evolution including branching, dynamically triggered slip and
back-propagation on different fault segments. Due to plastic
yielding between branching segments, part of the seismic energy
is absorbed by plastic processes, which affects in turn the on-
fault source dynamics. In comparison to an elastic scenario, off-
WELCOME
2016 SCEC Annual Meeting page 253
fault plasticity clearly reduces the peak slip rate on the whole fault
up to 30 %. The elastic and plastic scenarios exhibit distinct
spatio-temporal rupture transfers between fault segments, which
also alters the final slip distribution. Our results indicate that off-
fault plasticity plays an important role when assessing potential
seismic risks on complex fault geometries where parts of the fault
might be triggered dynamically.
Scaling of finite-source parameters at Parkfield,
California, Katie E Wooddell, Doug S Dreger, Taka'aki Taira,
Robert M Nadeau, and Luca Matagnini (Poster Presentation 064)
We determine finite-source slip models for Parkfield earthquakes
ranging in magnitude from M1.8 to M4.1 by inverting relative
moment rate functions (RMRF) obtained from empirical Green’s
function (eGF) deconvolution. The method is based on Mori and
Hartzell (1990). We then use the derived slip models to estimate
the stress change following the method of Ripperger and Mai
(2004), and compare with published results for the 2004 Parkfield
mainshock. These results provide an update to the study
presented at SSA in April 2016, and they show that many
microearthquakes display a level of complexity comparable to
slip models of larger earthquakes. We will present the results in
terms of the scaling of peak slip, average slip, and rupture area,
as well as the peak and average stress drops inferred from the
finite-source based stress change calculations. The updated
results indicate that on average the events behave self-similarly,
and the stress drops are consistent with corner frequency based
estimates of Imanishi and Ellsworth (2006), Allmann and Shearer
(2007), and Abercrombie (2014). On the other hand, the peak
slip and stress drops seem to follow the relationship proposed by
Nadeau and Johnson (1998) in which the peak stress drop
increases for decreasing scalar moment.
Stress Drop and Source Scaling of Recent Earthquake
Sequences across the United States, Qimin Wu, and Martin
C Chapman (Poster Presentation 241)
Knowledge of source scaling and variation in stress drop are
essential to a better understanding of source physics and strong
ground-motion prediction. Due to the large uncertainties in stress
drop measurements, many previous studies have found results
showing both self-similar and non-self-similar scaling
relationships between large and small earthquakes. Moreover,
the validity of the recently proposed idea that induced
earthquakes in the central US have low stress drops needs to be
tested in a reliable and systematic way.
We investigate the stress drop and source scaling of four recent
earthquake sequences across the US: the 2011 Mw 5.7 Mineral,
Virginia earthquake, the 2011 Mw 5.6 Prague, Oklahoma
earthquake, the 2014 Mw 6.0 South Napa, California earthquake
and the 2016 Mw 5.1 Fairview, Oklahoma earthquake. We apply
the coda spectral ratio method, which enables us to separate the
source effect from the path propagation and site response
effects, to obtain reliable estimates of corner frequencies and
stress drops. We demonstrate that the use of the coda spectral
ratio method can produce more stable estimates than the use of
the direct S-waves in the conventional empirical Green’s function
(EGF) or spectral ratio method. We carefully examine the decay
of coda envelopes between selected event pairs, and the corner
frequencies and stress drops are estimated by modeling the
observed source spectral ratio with the omega-squared source
spectral model. Our results shed light on the yet un-resolved
question whether potentially induced earthquakes are different
from tectonic earthquakes in terms of stress drop and source
scaling across the US.
Comparison of GPS Strain Rate Computing Methods and
Their Application, Yanqiang Wu (Poster Presentation 143)
Using modeled and simulated data for comparison of several
methods to compute GPS strain rate fields in terms of their
precision and robustness reveals that least-squares collocation
is superior. Large scale (75°E–135°E and 20°N–50°N) analyses
of 1° grid sampling data and decimated 50% data by resampling
(then erasing data in two 5°×10° region) reveal that the Delaunay
method has poor performance and that the other three methods
show high accuracy. The correlation coefficients between
theoretical results and calculated results obtained with different
errors in input data show that the order in terms of robustness,
from good to bad, is least-squares collocation, spherical
harmonics, multi-surface function, and the Delaunay method.
The influence of data sparseness on different methods shows
that least-squares collocation is better than spherical harmonics
and multi-surface function when sample data are distributed from
a 2° grid to a 1° grid. Strain rate results obtained for the Chinese
mainland using GPS data from 1999–2004 show that the
spherical harmonics method has edge effects and that its value
and range increase concomitantly with increased sparseness.
The multi-surface function method shows non-steady-state
characteristics; the errors of results increase concomitantly with
increased sparseness. The least-squares collocation method
shows steady characteristics. The errors of results show no
significant increase even though 50% of input data are decimated
by resampling. The spherical harmonics and multi-surface
function methods are affected by the geometric distribution of
input data, but the least-squares collocation method is not.
We used GPS velocities from approximately 700 stations in
western China to study the crustal deformation before the
Wenchuan MS 8.0 earthquake. The GPS strain rate in the E-W
direction in the Qinghai-Tibet block shows that extensional
deformation was dominant in the western region of the block
(west of 92° E), while compressive deformation predominated in
the eastern region of the block (from 92° E to 100° E). On a
regional scale, the hypocentral region of the Wenchuan
earthquake was located at the edge of an intense compression
deformation zone of about 1.9×10-8/yr in an east–west direction.
The characteristic deformation in the seismogenic fault was
compressive with a dextral component. The compression
deformation rate was greater in the fault’s western region than in
its eastern region, and the strain accumulation was very slow on
the fault scale.
The principal strain rate, derived from GPS velocities from the
Tibetan Plateau and the India Plate, shows that the Himalayan
tectonic belt exhibits compression deformation in a NE-NS-NE
direction from the west to the east. The GPS velocity profiles
reflect that the distribution of strain accumulation is uneven: there
is a 17.1 mm/yr compressive deformation distributed over 400
km along 85°E longitude, a 20.9~22.2 mm/yr compressive
deformation dispersed across 400~500 km along 79°E longitude,
and a 15.3~16.9 mm/yr compressive deformation spread across
500~600 km along 91°E longitude. The MW 7.8 Gorkha
WELCOME
2016 SCEC Annual Meeting page 254
earthquake occurred at the edge of an intense compression
deformation zone of about 6.0×10-8/yr in a north–south direction,
and an about 90% compressive strain is absorbed in the 300 km
region near the Main Frontal Thrust (MFT).
Using Well Water Level To Measure Volumetric Strain
Considering The Dissipation Effect, Yuqing Xie, Yonghong
Zhao, and Lingsen Meng (Poster Presentation 019)
Water-levels in confined well can be used as a strain-meter.
Periodic amplitude fluctuations of the water-level generally reflect
the strain change induced by Earth’s tide. The measured value
of well water-levels can be compared with the theoretical
volumetric tidal strain at various locations on earth. The
amplitude and phase delay between the tidal strain and water-
levels can be used to infer the hydrogeological parameters of
water-bearing rock. Traditional method analyzes amplitude and
phase delay in separate manners. In this study, a deconvolution
method considering both the amplitude and phase delay is
explored to remove the response of the well-aquifer system from
the volumetric strains. A numerical simulation is conducted to
demonstrate the feasibility of this method. We establish an ideal
2D confined aquifer model using poroelastic module in COMSOL
Multiphysics 5.1, a finite element software. We calculate the
response curve of water-levels under a step and periodical
changes of burden. Results show when the response curve is
known, the variation of the burden can be obtained by
deconvoluting response curve from the well water-levels.
Likewise, when the variation of burden is given, the response
curve can be also computed by deconvolving the loading curve
from the well water-levels. We apply the deconvolution method
to analyze the water-levels record of Huili Sichuan-06 well,
located at Sichuan Province, southwestern China. The response
curve is obtained using the predictive deconvolution filtering
method that deconvolves the theoretical value of volumetric tidal
strains from the measured value of well water level. Our results
show the response curve is similar to the simulation results of
ideal 2D model. Furthermore, the variation of additional
volumetric strain caused by regional stress field can be obtained
when deconvolving the response curve from measured value of
well water-levels after removing the effect of earth tide,
barometric pressure, etc. The response curve can be also used
to infer the sensitivity of well water level to the dissipation effect.
strain variation and dissipation law of abnormity of well water
level. It contains information of the well-aquifer system, including
hydrogeology properties that are hard to measure directly. Our
response curve also indicates that the dissipation process of
abnormity needs to be accounted for to accurately infer the
variation of volumetric strain caused by the regional stress field.
Tectonic and Anthropogenic Deformation Surrounding
the Cerro Prieto Geothermal Field from Sentinel-1
Interferometry, Xiaohua Xu (Poster Presentation 168)
The Cerro Prieto Geothermal Field (CPFG) lies at the step-over
between the Imperial and Cerrro Prieto Faults in northern Baja
California. While this is the most tectonically active section of the
San Andreas Fault system, the spatial and temporal deformation
in the area is poorly resolved by the sparse GPS coverage.
Moreover interferograms spanning more than a few months are
decorrelated due to the extensive agricultural activity in the
region. Here we investigate the use of frequent, short temporal
baseline interferograms offered by the new Sentinel-1 satellite to
recover two components of deformation time series across these
faults. The Sentinel-1 satellite uses a new wide-swath mode
called Terrain Observation with Progressive Scans (TOPS) to
acquire complete spatial coverage on a 12-day or 24-day
cadence from two look directions [Meta et al., 2010]. While this
new technique enables frequent acquisitions, it brings new
challenges to InSAR data processing in alignment and co-
registration [Prats-Iraola et al., 2012]. Following other groups, we
have developed a purely geometric approach to image alignment
that achieves better than 1/200 pixel alignment needed for
accurate phase recovery. One significant advantage of pure
geometric alignment is that phase correlation between SAR
images is not needed. Therefore long time span deformation can
be accurately constructed from a sum over short time span
interferograms because atmospheric and orbital errors of the
intervening SAR images cancel. In practice we construct InSAR
time series using a coherence-based SBAS method [Tong and
Schmidt, 2016] with atmospheric corrections by means of
common-point stacking [Tymofyeyeva and Fialko 2015].
Preliminary results, clearly resolve the spatial and temporal
subsidence at CPGF. The maximum subsidence rate of 150
mm/yr, due to extraction of geothermal fluids and heat,
dominates the expected 40 mm/yr deformation across the
proximal ends of the Imperial and Cerro Prieto Faults. We
combine a two-year time series of ascending (34) and
descending (42) Sentinel-1 images to map the details of the fault-
parallel deformation away from the CPFG.
Investigate fault zone hydrogeologic architectures by
using water level tidal and barometric response, Lian Xue,
Emily E Brodsky, Vincent Allegre, Patrick M Fulton, L. Parker L
Beth, and John A Cherry (Poster Presentation 105)
Fault zone hydrogeologic architecture is critical to faulting
processes; however, they are not well understood and difficult to
measure in situ. Water levels inside conventional water wells can
tap an artesian aquifer response to pressure head disturbances
caused by the Earth tides and surface atmospheric loading. The
fluctuation of water levels can measure the hydrogeologic
properties of the formation surrounding these wells. Specifically,
the amplitude of water level oscillation is mainly determined by
formation specific storage, and the phase shift between the water
level oscillation and the pressure head disturbance is mainly
determined by formation permeability. Utilization of wells at
diffident distances to faults is able to map fault zone
hydrogeologic architectures. We investigate fault zone
hydrogeologic architectures in two locations: Logan Quarry and
Southern Simi Valley, by applying water level tidal response. The
San Andreas Fault near Logan Quarry has a surprising uniform
diffusivity structure, while with higher specific storage and larger
permeability in a localized zone near the fault (within 40 m of the
fault). This change of properties might be related to the fault zone
fracture distribution. The faults in the southern Simi Valley also
have a uniform diffusivity structure but no significant fault-guided
permeability or compliance structures. Such homogenous by
fault zone damage is possible in a region of multiple strands and
copious secondary faulting. The observed uniform diffusivity
structure may suggest that the permeability contrast might not
efficiently trap fluids during the interseismic period. Both of these
two sites have a diffusivity of 10-2 m2/s which is also comparable
WELCOME
2016 SCEC Annual Meeting page 255
to the post-earthquake hydraulic diffusivity measured on the
Wenchuan Earthquake Fault. These two studies hint that
hydraulic diffusivity may evolve to a narrow range of values 10-2
m2/s in fault zones.
Comprehensive Detection and Relocation of the 2010
Mw7.2 El Mayor-Cucapah Foreshock Sequence,
Dongdong Yao, Zhigang Peng, Xiaofeng Meng, Xiaowei Chen,
Raul R Castro, and Sizuang Deng (Poster Presentation 216)
Large earthquakes are sometimes preceded by foreshocks, and
the exact relationship between foreshocks and mainshock
nucleation is still unclear. Here we present a comprehensive
foreshock detection of the 2010 Mw7.2 El Mayor-Cucapah,
Mexico, earthquake, using continuous waveforms recorded by
the South California Seismic Network (SCSN). Specifically,
continuous waveforms from SCSN stations within 150 km
epicentral distances were first downloaded. We use ~70
foreshocks listed in SCSN catalog [Hauksson et al., 2010] as
templates, and scan through continuous recordings starting from
20 days before up to the mainshock. Currently, ~400 events are
detected and we find nearly continuous episodes of foreshock
sequence, with the most intensive activity within 42 hours prior to
the mainshock. We also identify foreshocks from the triggered-
mode recordings by the Red Sísmica del Noroeste de México
(RESNOM) network. Even though there were no continuous
waveforms for detection, they can be used to verify the existence
of newly detected events and provide additional phase arrivals at
nearby stations. By combining phase information from RESNOM
and SCSN stations, we further perform HypoDD relocations for
all ~400 foreshocks. By closely examining spatio-temporal
evolutions of newly detected foreshock sequence, we hope to
better understand the nucleation process of the 2010 Mw7.2 El
Mayor-Cucapah mainshock.
Strike-slip Faulting Energy Release, Lingling Ye, Hiroo
Kanamori, Thorne Lay, and Jean-Philippe Avouac (Poster
Presentation 243)
Teleseismic recordings of P waves from 26 large (mostly, MW ≥
7.5) strike-slip earthquakes from 1990- 2015 have been analyzed
to determine their radiated energy, ER, exploring the stability of
the estimates with respect to the faulting geometry and directivity.
The radiated energy estimates demonstrate that moment-scaled
radiated energy ER/M0, for strike-slip events is systematically
higher than found for large megathrust events. Finite-fault
inversions for several of the events have been performed to
resolve the spatial extent of faulting and to calculate the stress
drops, supplementing published estimates for other events. The
effects of rupture directivity (deviation from point-source radiation
patterns) for the non-uniform slip models have been assessed for
several large ruptures by examining azimuthal patterns of
individual station ER estimates. The source directivity effect on
single-station ER measurements is generally relatively weak, but
systematic variation as a function of directivity parameter can be
detected for several events. A factor of ~2-3 azimuthal trend is
observed for most supershear events, and bias of the average
value can be avoided by azimuthal sampling. This effect appears
modest, but one must consider the narrow cone of teleseismic P
wave paths from the source and their proximity to the nodal
planes of strike-slip focal mechanisms. There are also
complexities in each rupture and directivity effects are not the
extreme end-member cases expected for simple uniform
unilateral ruptures. In addition, we estimate the radiated energy
for moderate earthquakes in Japan and South California with
dense regional seismic network to evaluate the uncertainty in
radiated energy measurements.
ETAS 2.x: The 2015 Nepal aftershock forecast, a Global
ETAS Code, and ETAS based Ground Motion
Forecastint, Mark R Yoder, John M Wilson, John B Rundle, and
Margaret T Glasscoe (Poster Presentation 303)
In this work, we present improvements, extensions, and
refinements to an ETAS type aftershock seismicity model,
originally developed and published by Yoder et al. (2015, 2014
online). In our previous work, we show that near-field aftershock
rate-densities can be accurately estimated based on well known
earthquake scaling relations and geometric constraints. Of
recent note, this model was used to produce the aftershock
forecast that accurately predicted the locations of the m=7.3 and
m=6.3 aftershocks following the 2015 m=7.8 Gorkha, Nepal
mainshock event. Aftershock forecasts based on this model have
also been posted for the large, m=7.8 event in Ecuador and the
m=7 Kyushu, Japan earthquakes -- both in the Spring of 2016,
as well as for several simulated disaster response exercises,
including the FEMA 2016 "Cascadia Rising" scenario. In this
work, we discuss the application of this model during the Nepal
response, and we perform ensemble testing that shows that the
model produces significant information gain and forecast skill.
We have also developed an improved version of the code, which
is freely available for use and collaborative development via
GitHub. The new code is more streamlined and accessible to
developers and practitioners. Perhaps more importantly, we
have reorganized the workflow to enable parallel processing, and
we have introduced a spatial indexing scheme that, by limiting
the effective range of small earthquakes, facilitates very large,
even global forecasts. We further extend the model by integrating
it with ground motion prediction equations (GMPE). We invert the
ETAS forecast to produce an array of source events, which we
plug into standard point-source GMPE, to produce a map of
expected threshold exceedance rates.
Preliminary report on site characterization using
noninvasive single- and multi-station methods at southern
California seismic stations, Alan Yong, Antony Martin,
Jennifer Pfau, Devin McPhillips, Marcos Alvarez, Scott S Lydeen,
Fiona Clerc, and Nolan Leue (Poster Presentation 226)
In-situ measurements of shear-wave velocity (Vs) are used
commonly to evaluate seismic response at earthquake
monitoring station and project sites. Vs30, the time-averaged Vs
in the upper 30 m, is a common parameter used to capture
seismic site response and is used in almost all modern ground
motion prediction equations. Traditional invasive downhole
methods directly measure Vs; however, these methods are often
cost- and/or environmentally-prohibitive and their results do not
always reflect the lateral variability of seismic conditions beyond
the immediate vicinity of the test site. In comparison, noninvasive
methods record active- or passive-source data consisting of
surface or body waves and are less prohibitive to use. Moreover,
these methods use multiple horizontally-spaced surface
receivers (multi-station array), thus, lateral variability beneath the
array is accounted for in their results. Most noninvasive methods,
WELCOME
2016 SCEC Annual Meeting page 256
however, indirectly measure Vs, and thus have inherent
uncertainties. We have used a suite of noninvasive methods at
six stations in southern California. We record microseisms using
standalone single-stations, located at the end- and mid-points of
the measurement array, and over the same period, we also
collect records from the seismic station. Using both single- and
seismic-station records, we calculate the horizontal-to-vertical-
spectra-ratios (HVSR), resonance frequency, and power spectral
density to study site characteristics, including noise levels. For
soil sites, we generally find insignificant lateral variability in
subsurface conditions beneath our multi-station arrays by
matching similar spectral peaks and frequencies in the three
HVSR records; for rock sites, the magnitudes of the HVSR values
are not as discernible. While we find general agreement in Vs30
computed using a variety of methods at each site, preliminary
results for low-noise sites using standalone passive methods
have large uncertainty in their computed Vs30 values.
Earthquakes Induced by Hydraulic Fracturing and
Wastewater Injection in Guy-Greenbrier, Arkansas, Clara
E Yoon, Yihe Huang, William L Ellsworth, and Gregory C Beroza
(Poster Presentation 194)
The Guy-Greenbrier, Arkansas, earthquake sequence, which
occurred from July 2010 through October 2011, was potentially
induced by injection of wastewater from nearby hydraulic
fracturing operations into disposal wells (Horton, 2012). To gain
insight into the initial stages of this earthquake sequence, we
detected and located earthquakes during a three month time
period from 2010-06-01 to 2010-08-31, spanning a time interval
before and after the beginning of the sequence in July 2010. We
then examined spatial and temporal correlations between
seismicity, wastewater injection, and hydraulic fracture
stimulation of production wells.
Although the Arkansas seismic network is sparse, the Fingerprint
And Similarity Thresholding (FAST) method (Yoon et al., 2015)
enabled comprehensive detection of ~14,000 low-magnitude
earthquakes ranging from M ~ -1 to 1.8 in continuous seismic
data from a single 3-component station WHAR. We then located
the largest 756 (M > 0) earthquakes with data from two additional
stations, ARK1 and ARK2, in a small local network, using the
velocity model from Ogwari et al. (2016).
The majority of these earthquakes are located at the north end of
the Guy-Greenbrier Fault, a previously unknown fault that would
later be illuminated by hundreds of thousands of earthquakes in
2010-2011. By comparing the stimulation and disposal history of
wells in the area with precise earthquake locations, we find that
many of the earthquakes are closely correlated in space and time
with either hydraulic fracturing operations or wastewater
injection. We also observe several distinct clusters of seismicity
off the Guy-Greenbrier Fault that closely follow hydraulic
fracturing stimulation at a nearby production well.
Our retrospective analysis indicates that high accuracy
earthquake locations with low magnitude detection thresholds
can provide new insights into sources of potentially induced
seismicity.
Estimation of physical properties of a rock sample based
on a laboratory transmitted wave experiment and 3D
numerical simulations, Nana Yoshimitsu, and Takashi
Furumura (Poster Presentation 036)
Elastic wave transmission during a rock compression test is an
efficient way to estimate the characteristics of a sample. Full
utilization of transmitted waveform would lead much better
understanding of the sample interior, in addition to the utilization
of initial motion. To clarify the wave behavior of multiple
reflections and conversions in a finite sample, we performed a
3D numerical simulation of wave propagation in a homogeneous
laboratory-scaled sample, and obtained good agreement of the
waveforms between the simulation and observation. In this study,
we introduce a low velocity zone to a model as a fractured zone
with time-dependent velocity variation, and compare the
waveforms obtained by the simulation and laboratory
compression test.
We calculated 3D finite difference method simulation of seismic
wave propagation with 4th- and 2nd- order accuracy in space and
time referring to the tri-axial compression test of Imahori et al.
(2015). The base model had homogeneous cylindrical structure
with 100 mm height and 50 mm diameter which grid spacing was
54 and 60 micrometers in horizontal and vertical directions,
respectively. Single force was added on the surface of the model
to mimic the movement of the transducer. We used estimated P-
and S-wave velocity (Vp = 4100 – 5400 m/s; Vs = 2200 – 3100
m/s), rigidity = 2.7 g/cm3, respectively. A fault which tilted 30
degrees from the axis of the cylinder with 0.3 mm width was
modeled by referring the X-ray CT image of the recovered granite
sample after the tri-axial compression test. The fault zone was
regarded as a thin low velocity zone here, where we assumed Vp
= 2.7 km/s, Vs = 1.55 km/s, and rigidity of 1.35 g/cm3,
respectively. Band-pass filter of 200-400 kHz was applied to the
calculated waveforms.
Each phases in the later part of the observed transmitted waves
(Imahori et al., 2015) showed different behavior as the loading
progress; some phases disappeared and some newly appeared.
We applied two pairs of the velocity model to reproduce the
waveforms obtained in the earlier stage and the later stage of the
loading in the experiment, and both case showed good
agreement of the waveforms between the simulation and
observation. Newly appeared characteristic phase in the later
stage of the loading was well reproduced by the simulation. This
agreement suggests the possibility of the interpretation and
utilization of the later phases of the transmitted wave.
New Holocene slip-rate sites along the Mojave San
Andreas Fault near Palmdale, CA, Elaine K Young, Eric S
Cowgill, and Katherine M Scharer (Poster Presentation 118)
Geologic and geodetic slip rates for the Mojave segment of the
San Andreas fault (MSAF) appear to be discrepant: the
Quaternary geologic rate is as high as 37 mm/yr [1] while the
geodetic slip rate is as low as ~15 mm/yr [2]. To better
characterize the Holocene-average rate for the MSAF, we
investigated several offset landforms near Palmdale, CA using
detailed surficial mapping and 14C analyses of charcoal
collected from hand-dug excavations. The first site, “X-12”,
preserves two offset markers sourced from a north-flowing
catchment that is south of the fault. The first marker is a west-
facing terrace riser that was cut into older alluvium (Qoa) and
offset ~75 m. The base of the riser is abutted by an alluvial fan
WELCOME
2016 SCEC Annual Meeting page 257
(Qf) deposited against the riser. Progressive offset of the Qoa
riser and Qf fan generated a north-facing fault-scarp that then
eroded to deposit a small fan (Qff) on top of Qf north of the fault
scarp. The second offset is a beheaded channel incised into Qf
north of the fault that is separated 50 m from the source channel
to the south. To date these landforms we opened five
excavations at the site, with emphasis on the beheaded channel.
Dates from within the Qf fan south/upstream of the fault and from
abandoned bedload in the beheaded channel north/downstream
of the fault provide maximum and minimum bounds on the age
of incision of the beheaded channel of ~1500 calBP and ~600
calBP, respectively. Combining the 50 m offset with a maximum
depositional age implies a slip rate of ≥ 33 mm/yr since the Qf
fan was incised. Future work will date 1) post-abandonment
deposits within the beheaded channel to obtain a minimum age
for the beheaded channel and 2) Qff to provide an upper bound
on the slip rate for the 75 m offset. At a second site, Ranch
Center, a north-flowing stream cut across a shutter ridge north of
the fault and was then offset and deflected ~80 m before cutting
a new channel across the fault. Future work here will date the
base of alluvial fans deposited by the abandoned and active
channels to obtain maximum and minimum ages for the ~80 m
offset, respectively. Rates from these two sites will help to better
characterize the geologic Holocene slip rate on the MSAF. Future
work will use rates from this study, and others, to explain the
difference between longer-term geologic and more recent
geodetic slip rates on the MSAF.
[1] Matmon et al., 2005, GSAB. v. 117 p. 795
[2] Becker et al., 2005, Geoph.. J. Int., v. 160 p. 634
Constraints on residual topography and crustal properties
in the western US from virtual deep seismic sounding,
Chunquan Yu, Wang-Ping Chen, and Rob van der Hilst (Poster
Presentation 204)
We use virtual deep seismic sounding (VDSS) and data from
~1,000 broadband seismic stations to provide high-resolution
estimates of crustal structure in the western Cordillera of the
United States (US). The most robust result is the geographic
distribution of residual topography (that is, the difference
between observed elevation and that expected from crustal
buoyancy alone) and, by implication, thermal or petrologic
anomalies in the mantle. Overall, residual topography of the
western US Cordillera varies considerably; with contrasts of up
to about 3 km across distances of 200 km or less. High residual
topography, indicating large mantle effects, is evident along the
periphery of the Colorado Plateau and the surroundings of the
Great Basin. In contrast, the central Colorado Plateau and the
Wyoming Basin show low residual topography, close to what is
expected of a geologically stable lithosphere. Overall, in regions
to the east of the Wasatch hinge line (the eastern limit of
significant extension in the North American cratonic basement)
patterns of high residual topography and anomalies of low
seismic wave-speeds in the upper mantle are similar, suggestive
of a common, thermal origin. In contrast, such a similarity is
absent in regions to the west of the hinge line, suggesting
substantial effects of petrological heterogeneities in the mantle.
Finally, joint analyses of VDSS and conventional receiver
functions reveal a wide range of crustal P-wave speeds, locally
as high as 6.7 km/s, perhaps indicating magmatic modification of
the crust.
Products and Services Available from the Southern
California Earthquake Data Center (SCEDC) and the
Southern California Seismic Network (SCSN), Ellen Yu,
Prabha Acharya, Aparna Bhaskaran, Shang-Lin Chen, Jennifer
Andrews, Valerie Thomas, Egill Hauksson, and Robert W
Clayton (Poster Presentation 198)
The SCEDC archives continuous data from 9429 data channels
and 513 SCSN recorded stations. The SCEDC processes and
archives an average of 16,000 earthquakes each year. The
SCEDC provides public access to these earthquake parametric
and waveform data through its website scedc.caltech.edu and
through web services and client applications such as STP. This
poster will describe the most significant developments at the
SCEDC in the past year.
In addition to the real time GPS displacement waveforms now
archived at SCEDC, the SCEDC is now archiving
seismogeodetic waveforms displacement and velocity
waveforms produced by the California Real Time Network. These
waveforms are computed by combining the 1 sps real time GPS
solutions with 100 Hz accelerometer data.
The SCEDC data holdings now include a double difference
catalog (Hauksson et. al 2011) spanning 1981 through 2015
available via STP, and a focal mechanism catalog (Yang et al.
2011).
In an effort to assist researchers accessing catalogs from
multiple seismic networks, the SCEDC has entered its
earthquake parametric catalog into the ANSS Common Catalog
(ComCat). Origin, phase, and magnitude information have been
loaded.
The SCEDC has implemented Continuous Wave Buffer (CWB)
to manage its waveform archive. This software was developed
and currently in use at NEIC. CWB has simplified and
streamlined waveform archival, which allows the SCEDC to
maximize the completeness of its waveform archives as well as
making continuous data available within seconds of real time.
The SCEDC now provides web services to access its holdings.
Event Parametric Data (FDSN Compliant):
http://service.scedc.caltech.edu/fdsnws/event/1/
Station Metadata (FDSN Compliant):
http://service.scedc.caltech.edu/fdsnws/station/1/
Waveforms (FDSN Compliant):
http://service.scedc.caltech.edu/fdsnws/dataselect/1/
Event Windowed Waveforms, phases:
http://service.scedc.caltech.edu/webstp/
Kinematic Rupture Process of the 2016 Mw 7.1
Kumamoto Earthquake Sequence, Han Yue, Mark Simons,
Cunren Liang, Heresh Fattahi, Eric J Fielding, Hiroo Kanamori,
Donald V Helmberger, Linjun Zhu, Michael Sylvain, and Jean-
Philippe Avouac (Poster Presentation 076)
On April 16th, 2016, the Mw 7.1 Kumamoto earthquake ruptured
a portion of the Futagawa fault on the Kyushu island. This event
WELCOME
2016 SCEC Annual Meeting page 258
was preceded by two (Mw 6.0 and 6.2) foreshocks happened two
days earlier. We investigate the kinematic rupture process of the
Kumamoto earthquake sequence by integrated analysis of
synthetic aperture radar (SAR) interferometry and azimuth-offset
images and regional strong motion records. From the SAR
azimuth offset images, which covers both the foreshocks and
main shock displacement fields, we identify a “bifurcated” fault
trace. The main rupture presents approximately 40km lateral
extent striking 220 to the southwest. A 10 km long fault branch
connected to the southern portion of the main rupture is also
identified. Guided by the surface fault trace and aftershock
distribution at ~16 km depth, we parameterize the main rupture
plane with a curved fault interface. The dip angle of the main fault
plane change gradually from 85 (south) to 75 (north). Static GPS,
SAR interferometry and azimuth-offset images and near field
strong motion records of 7 K-net and KiK-net stations are
adopted jointly in both linear and Bayesian inversions. For a
portion of SAR images record both foreshocks and the main
shock deformation fields, we parameterize the source time
function as foreshock (static) and main-shock (kinematic)
ruptures and specify their contribution to each dataset. This
enables us to invert for the foreshock and main shock ruptures
simultaneously in one inversion work. A unilateral rupture
process is resolved in both inversions, with the main shock
rupture extends for ~30 km from the hypocenter to the northeast.
Peak slip of approximately 6 m is resolved at ~5 km depth which
locates ~25 km from the hypocenter. Significant shallow slip
deficit is resolved at shallow depth with maximum surface rupture
of ~2 m. We also resolved significant normal fault slip
components, which is consistent with the normal faulting
component revealed by GCMT solutions. Both inversion results
revealed averaged rupture velocity (Vr) of ~ 2.9 km/s, while the
Bayesian inversion resolved several local areas with high rupture
velocity. Several sub-parallel off-fault fractures are identified from
the InSAR images near the northern end of the main rupture
area. These off-fault fractures form a conjugate angle with the
main fault plane. Those fractures may be activated by the strong
shearing stress of the dynamic waves, while their normal faulting
mechanism is consistent with the dilatational back-ground stress
field of this volcanic area. Both main shock and aftershocks
presents a combination of normal and strike slip components,
which indicates the regional stress field is controlled by both the
tectonic shearing and the volcano introduced dilatation.
Earthquake Declustering via a Nearest-Neighbor
Approach, Ilya Zaliapin, and Yehuda Ben-Zion (Poster
Presentation 310)
We propose a new method for earthquake declustering based on
nearest-neighbor analysis of earthquakes in space-time-
magnitude domain. The nearest-neighbor approach was recently
applied to a variety of seismological problems that validate the
general utility of the technique and reveal the existence of several
different robust types of earthquake clusters. Notably, it was
demonstrated that clustering associated with the largest
earthquakes is statistically different from that of small-to-medium
events. In particular, the characteristic bimodality of the nearest-
neighbor distances that helps separating clustered and
background events is often violated after the largest earthquakes
in their vicinity, which is dominated by triggered events. This
prevents using a simple threshold between the two modes of the
nearest-neighbor distance distribution for declustering. The
current study resolves this problem hence extending the nearest-
neighbor approach to the problem of earthquake declustering.
The proposed technique is applied to seismicity of different areas
in California (San Jacinto, Coso, Salton Sea, Parkfield, Ventura,
Mojave, etc.), as well as to the global seismicity, to demonstrate
its stability and efficiency in treating various clustering types. The
results are compared with those of alternative declustering
methods.
Correlation between Strain Rate and Seismicity and Its
Implication for Earthquake Rupture Potential in California
and Nevada, Yuehua Zeng, Mark D Petersen, and Zheng-Kang
Shen (Poster Presentation 307)
Seismicity and strain-rate patterns indicate a region of increased
hazard for large earthquakes across a broad region of high strain
rates found mostly in California and western Nevada. These
results are based on over three decades of GPS measurements
and an earthquake catalog that spans the past century. From the
geodetic data we develop strain-rate maps and correlate these
with seismicity data. These correlations allow us to identify
patterns of loading and unloading. Patterns of activity change
with time, evolving through a (1) accumulation phase, (2)
localization phase, and (3) rupture phase that have been
observed in rock mechanics studies and dynamic simulations of
fault development over many earthquake cycles. Following the
approach of Helmstetter et al. (2007) and Shen et al. (2007), we
classify strain rates as high or low and correlate these levels with
changes in regional seismicity. We found that many of the
moderate to large (M≥6.5) earthquakes are collocated with the
regions of highest secular strain rates (e.g., the San Andreas
fault system) whereas the smaller earthquakes are often spread
out across regions of both high and low strain. For earthquakes
with magnitude above M4.0 (completeness level since 1933), we
identify clear spatial and temporal changes. For example, from
1933 to the mid 1990’s smaller magnitude seismicity occurred in
both high and low strain rate regions (accumulation phase). From
the mid 1990’s to 2013 this pattern changed and seismicity was
more concentrated within the high strain rate areas (localization
phase). The large-scale spatial and temporal changes in
seismicity pattern may be consistent with future increased activity
along faults in the active strain region and weaker activity in the
surrounding areas. This pattern is similar to the precursory
earthquake model defined as the Mogi doughnut (Mogi, 1969).
This increased rate of earthquakes in high strain rate areas and
decreased rate of earthquakes in low strain rate areas has many
implications for future earthquakes in California and Nevada.
Friction and stress drop controlled by fault roughness and
strength heterogeneity, Olaf Zielke, Martin Galis, and Paul M
Mai (Poster Presentation 052)
An earthquake's stress drop is a fundamental quantity of the
rupture process, intimately related to the decrease of frictional
resistance to sliding during rupture evolution. A number of recent
high-speed laboratory friction experiments have shown a nearly
complete breakdown in frictional resistance when coseismic slip
velocities are reached. Such large changes in friction coefficient
imply large stress drops that exceed those reported for natural
earthquakes by an order of magnitude or more. Laboratory
WELCOME
2016 SCEC Annual Meeting page 259
friction experiments and seismologic observations seem to
contradict each other.
Using large-scale numerical simulations, we address this
discrepancy between laboratory- and field-based stress drop
estimates. We show that it can be resolved when considering
fractal fault geometries and spatial heterogeneity of fault
strength. We find that stress drop estimates for natural
earthquakes are consistently too low because the underlying
analytical expressions assume a planar instead of a
geometrically complex source. Fault roughness serves as a bulk
frictional agent. To achieve compliance in laboratory- and field-
based stress drop estimates additionally requires a
heterogeneous distribution of fault strength: Laboratory friction
experiments represent characteristics of strong fault asperities
(stress drop estimates of high-strength regions) while field
observations provide space-time averaged estimates of stress
drop, containing asperities and non-asperities. The distribution of
strength asperities across a rupture surface has a profound
impact on an earthquake’s moment release. Unilateral ruptures
– with near-the-edge strength asperities – release less seismic
moment per stress drop than bilateral ruptures with well-centered
strength asperities.
Lidar mapping and luminescence dating reveal highly
variable latest Pleistocene-Holocene slip rates on the
Awatere fault at Saxton River, South Island, New
Zealand, Robert Zinke, James F Dolan, Russ J Van Dissen, Ed
J Rhodes, and Christopher P McGuire (Poster Presentation 100)
We use high-resolution lidar micro-topographic data and
luminescence dating to constrain incremental Holocene–late
Pleistocene slip rates at the well-known Saxton River site along
the Awatere fault, a primary dextral strike-slip fault in the
Marlborough Fault System, South Island, New Zealand. Field
work and mapping using lidar-derived topography yields revised
measurements of seven fault offsets recorded by fluvial terrace
risers, channels, and bedrock features. Improved dating of those
features is provided by stratigraphically informed analysis of >20
post-IR50-IRSL225 luminescence ages. Monte Carlo analysis is
used to derive robust determinations of incremental slip rates.
Our results show definitively that, far from being constant,
incremental Awatere fault slip rates (averaged over multiple
millennia and 10’s of m of slip) have varied by more than an order
of magnitude since latest Pleistocene time. These observations
have major implications for earthquake occurrence, plate
boundary lithosphere behavior, and probabilistic seismic hazard
assessment.
Floods, storms, and the identification of wave-dominated
deltas: Insights from ground-penetrating radar profiles of
the Oxnard Plain, Julie M Zurbuchen, and Alexander R Simms
(Poster Presentation 106)
Tsunamis often occur in tectonically active areas, where they
pose a significant risk to property and life along low-elevation
coastal regions. In California, tsunami studies have traditionally
focused on the Cascadia Subduction Zone in the north, however
historical records and recent models, including those for the
Ventura-Avenue Anticline-Pitas Point Thrust, have shown that
tsunamis may pose a significant threat to the southern California
coastline as well. A ground-penetrating radar (GPR) survey was
collected along the shoreline of the Oxnard Plain to look for
evidence of erosion caused by past tsunamis. 100 MHz, 200
MHz, and 500 MHz GPR profiles were collected at five state
beaches in the area penetrating depths of up to 6, 3, and 2
meters, respectively. Although several erosional surfaces were
identified, no evidence for tsunami erosion was found within the
GPR profiles. However, to date, the length of our record is
unknown. The identified erosional surfaces are interpreted as
storm erosion due to their frequency and occurrence in parts of
the coastal plain that formed over historical periods. The
prevalence of landward-dipping reflections near the Santa Clara
River delta may provide a stratigraphic signature of events
associated with large sediment output from the river (i.e large
inland storms) and a means of distinguishing wave-dominated
deltas from prograding clastic shorelines in the ancient rock
record.
WELCOME
2016 SCEC Annual Meeting page 260
Meeting Participants AAGAARD Brad, USGS 043
ABATCHEV Zagid, UCLA 254
ABERCROMBIE Rachel, Boston U 063,
222, 225
ABOLFATHIAN Niloufar, USC 013
ABRAHAMSON Norman, PG&E 292
ACHARYA Prabha, Caltech 198
ADAMS Mareike, UCSB 057, 240
ADAMS Paul, Aerospace Corp 085
ADLER Michaela, CSU Fullerton
AGNEW Duncan, UCSD 261
AGONCILLO Kristina Marie, CalPoly
Pomona
AGUIRRE Jorge, 234
AHDI Sean, UCLA
AHMAD Mohammad, 187
AIKEN John, GFZ Potsdam 312
AKCIZ Sinan, CSU Fullerton 128, 130
AKTUG Bahadir, 135
ALLAM Amir, U Utah 185, 199, 201, 213
ALLEGRE Vincent, UCSC 105
ALLEN Richard, UC Berkeley 249, 319
ALLISON Kali, Stanford 321
ALMOND Peter, 084
ALVARADO J. L., 031
ALVAREZ Karen, Cal Poly Pomona
ALVAREZ Marcos, New Mexico Tech 226
AMINZADEH Fred, USC 345
ANDERSON Greg, NSF
ANDERSON John, UNR 265, 281, 291
ANDREINI Jeremy, 116
ANDREWS Dudley Joe
ANDREWS Jennifer, 187, 198
ANGSTER Stephen, UNR 094
ARANHA Mario, 319
ARCE Adam, UCSB 057, 258
ARCHULETA Ralph, UCSB 289
ARGUS Donald, JPL 148
ARON Felipe, Catholic Univ of Chile 111,
325
ARROWSMITH Ramon, ASU 128, 130
ASIMAKI Domniki, Caltech 266
ASPIOTES Aris, USGS 145
ATKINSON Gail, Western U (Canada)
AUSTIN Kenneth, UNAVCO 144
AVOUAC Jean-Philippe, Caltech 076, 148,
243, 305
AYOUB Francois, Caltech 088
AYTUN Alkut, 135
AZEVEDO Steve, 346
AZIZZADEH-ROODPISH Shima, CERI
BABCOCK Jeff, UCSD 086
BACHHUBER Jeffrey, PG&E 084
BADEN Curtis, Stanford 111
BADER Michael, Tech U Munich 341
BAE Sung, QuakeCoRE 350
BAEZ Juan Carlos, 192
BAHADORI Alireza, Stony Brook 008, 165
BAKER Ben, 237
BAKER Jack, Stanford 296
BALLMANN Jason, SCEC/USC 327, 328,
334
BALTAY Annemarie, USGS 175, 278
BANDY Terryl, 133
BAO Han, UCLA 208
BARALL Michael, Invisible Software
BARBA Magali, CIRES UC Boulder 164
BARBERY Monica, TAMU 032
BARBOT Sylvain, EOS 148
BARBOUR Andrew, USGS 183, 278
BARCENA Hebert, CICESE
BARNETT Elizabeth, Shannon & Wilson
BARNHART William, U Iowa 133
BARRAGAN Katherine, CalPoly Pomona
BARRIENTOS Sergio, 192
BARSEGHIAN Derik, USGS 145
BARTEL Beth, UNAVCO
BARTH Nicolas, UCR 114
BASHIOUM Josh, Early Warning Labs
BASSET Andre, UNAVCO 144
BAUER Michael, 196
BAYLESS Jeff, AECOM / UC Davis 170,
292
BEAUDOIN Bruce, IRIS 346
BECKER Thorsten, UT Austin
BEELER N., USGS 029
BEHL Richard, CSULB 007
BEHN Mark, WHOI 339
BEHR Whitney, UT Austin 090, 127
BELL Katy Croff, 113
BELVOIR Sophia, SCEC USEIT 333
BEMIS Sean, U Kentucky 102
BEN-ZION Yehuda, USC 013, 120, 174,
185, 188, 201, 218, 219, 220, 229, 310
BENNETT Richard, U Arizona
BENOIT Margaret, NSF
BENT Morgan, USC 329
BENTHIEN Mark, SCEC/USC 327, 328,
331, 332, 333, 334
BENTZ John, UCSB 087
BERELSON Will, USC 311
BERG Elizabeth, U Utah 199
BERGEN Karianne, Stanford
BERNARDINO Melissa, UC Boulder
BEROZA Gregory, Stanford Tue 08:00,
186, 188, 193, 194, 273, 274
BETH L. Parker, 105
BEUTIN Thomas, GFZ Potsdam 313
BEYER Jennifer, UMass Amherst 058, 060
BHASKAR Ramya, 318
BHASKARAN Aparna, Caltech 198
BIASI Glenn, UNR 291, 297
BIEDA Allison, Group Delta
BIELAK Jacobo, CMU 272, 294
BIJELIC Nenad, Stanford Wed 08:30
BILHAM Roger, CIRES UC Boulder 116,
135
BIRD Peter, UCLA 009
BLANPIED Michael, USGS 024
BLETERY Quentin, U Oregon 070
BLEWITT Geoffrey, UNR Tue 10:30
BLISNIUK Kimberly, SJSU 117, 122
BOCK Yehuda, UCSD 153, 256
BODIN Paul, U Memphis 237
BODIN Thomas, 230
BOESE Maren, Caltech 192
BOETTCHER Margaret, U New Hampshire
053
BOLER Frances, UNAVCO Tue 10:30
BOLOGNA Alexander, USC
BOOKER Cecilia, NAVY
BORHARA Krishna, Utah State
BORMANN Jayne, UNR 109
BORSA Adrian, UCSD 165
BOSTOCK Michael, UBC (Canada) Tue
08:00
BOUÉ Pierre, Stanford 273
BOWIE Elliot, U Otago 290
BOZORGNIA Yousef, UC Berkeley
BRADBURY Kelly, Utah State 110
BRADLEY Brendon, QuakeCoRE 283,
284, 285, 286, 296, 350
BRANSCOMB Sally, ASU 128
BRANUM Dave, CGS
BRAZA Mary, UCLA
BRENNAN Michael, 113
BREUER Alex, Tech U Munich 336
BREUER Alexander, UCSD 340, 342
BRIGGS Rich, USGS 133
BROCHER Thomas, USGS 280
BRODSKY Emily, UCSC 034, 054, 105,
169, 202, 255
BROOKS Benjamin, USGS 075, 192
BROWN Kevin, SIO 037
BROWN Nathan, UCLA 117, 130
BRUDZINSKI Michael, UIUC 189
BRUHAT Lucile, Stanford 161
BRUNE James, UNR 291
BRUNE Richard, 291
BUCKLAND Kerry, 085
BUEHLER Janine, UCSD
BUGUEÑO Constanza, 325
BUNDS Michael, Utah State 116
BUNN Julian, Caltech 209
BURGETTE Reed, NM State Tue 10:30,
096, 099
BÜRGMANN Roland, UC Berkeley 012,
065, 157, 238
BURKHARDT Geneva, USC 334
BUTCHER Amber, CalPoly Pomona
BYDLON Samuel, Stanford 268
CACCAMISE Dana, NOAA
CADENA Ana
CAKIR Ziyadin, 135
CALLAGHAN Scott, SCEC/USC 002, 335,
337
CAMBRON JAZIEL, CICESE 073
WELCOME
2016 SCEC Annual Meeting page 261
CARLSON Jean, UCSB 039
CARLSON Kade, U Kentucky
CARPENTER Brett, U Oklahoma 025, 027
CARPENTER David, 252
CASEY Bridget, Virginia Tech 214
CASTILLO Christopher, Stanford 111, 113
CASTRO Raul, CICESE 216
CATCHINGS Rufus, USGS 005, 081, 176,
257, 262
CATO Kerry, CSUSB
CATTANIA Camilla, Stanford 044
CELEBI Mehmet, USGS
CEMBRANO José, 325
CERVANTES Rafael, ELAC 332
CHAN Joanne, USGS 257, 262
CHANG Emil, UCLA 079
CHAPMAN David, U Utah 011
CHAPMAN Martin, 241
CHAYTOR Jason, Oregon State 113
CHEN Min, Rice 246
CHEN Qianli, UIUC 182
CHEN Shang-Lin, Caltech 198
CHEN shi, CEA (China)
CHEN Tao, CEA (China) 080
CHEN Wang-Ping, 204
CHEN Xiaowei, U Oklahoma 027, 063,
184, 202, 216
CHENG Yifang, USC 184
CHERRY John, 105
CHESTER Frederick, TAMU 031, 032
CHESTER Judith, TAMU 032, 131
CHIOU Ray, NAVFAC EXWC
CHOURASIA Amit, SDSC/UCSD 347, 348
CHRISTENSEN Alan, City of Arvin 262
CHRISTOPHERSEN Annemarie, GNS
Science Tue 14:30
CHU Risheng, WHIGG 260
CHU Shanna, Stanford 046
CHUANG Lindsey, Tue 08:00
CLARE Richard, 350
CLARKE Parker, 132
CLAYTON Robert, Caltech 190, 198, 209
CLERC Fiona, 226
COCHRAN Elizabeth, USGS 187, 192,
197, 231
COFFEY Genevieve, Columbia 025
COHEN Hannah, CSUN
COLEMAN Dwight, 113
COLLETTINI Cristiano, 025
CONTRERAS-ZAZUETA Marcial, 295
COOKE Michele, UMass Amherst 015,
058, 059, 060
CORMIER Marie-Helene, U Missouri 113
COWGILL Eric, UC Davis 118
CRANE Thomas, CDC - DOGGR 101
CREMPIEN Jorge, UCSB 289
CRILEY Coyn, 257, 262
CROUSE C.B., AECOM Wed 09:00
CROWELL Brendan, U Washington 141,
153, 237
CUI Yifeng, SDSC 267, 272, 336, 337, 340,
342, 344, 347
CURRIE Brian, 189
DA SILVA LIMA Jadson, U Brasilia 333
DALRYMPLE William, UCR 252
DAVIS Craig, LADWP
DAVIS Paul, UCLA 177, 254
DAWSON Timothy, CGS
DAY Steven, SDSU 267, 344, 349
DE BARROS Louis, 070
DE CRISTOFARO Jason, USGS 223
DE GROOT Robert-Michael, USGS 231
DE PAOLA Nicola, Durham U
DE PASCALE Gregory, U Chile 084
DEFRISCO Michael, CGS
DEIERLEIN Gregory, Stanford Wed 08:30
DELBRIDGE Brent, UC Berkeley 065
DELEON Jonathan, 187
DEMINE Lauren, 252
DEMOREST John
DENG Sizuang, 216
DENNIS Kristen, SDSU
DENOLLE Marine, Harvard 273, 274
DESANTO John, UCSD 134
DESHON Heather, 024
DETERMAN Daniel, USGS 145
DEVECCHIO Duane, ASU
DEVRIES Phoebe, Harvard 018
DI ALESSANDRO Carola, GeoPentech
DI TORO Giulio, INGV 026
DICKINSON Haylee, 157
DIETERICH James, UCR 316, 322, 347
DITTMAN Stephen, 144
DOGRU Asli, 135
DOLAN James, USC 017, 088, 100, 101,
311
DONNELLAN Andrea, JPL 138, 152, 343
DONOVAN Jessica, RMS
DORSEY Matthew, SDSU
DOSHI Aadit, USC 334
DOUGHERTY Sara, USGS 181, 197
DOUILLY Roby, UCR 058
DREGER Doug, UC Berkeley 064
DRISCOLL Neal, UCSD 086, 107, 109
DU Ke, Stanford
DUAN Benchun, TAMU 048, 049, 050
DUBLANCHET Pierre, 068
DUDASH Stephanie, USGS 108
DUNBAR Sean, CA DWR
DUNCAN Christopher, 192
DUNHAM Eric, Stanford 045, 046, 072,
268, 320, 321
EBERHART-PHILLIPS Donna, UC Davis
021
EBRAHIMIAN Hamed, Caltech
ECHEVERRIA Henry, CalPoly Pomona
EDGINGTON Joshua, TAMU
ELBANNA Ahmed, UIUC 039, 040, 041,
042, 182
ELIZONDO Daniel, TAMU 131
ELLIS Susan, GNS Science 021
ELLSWORTH William, Stanford 053, 181,
193, 194, 217
EMBE Linus, Assoc Community
Awareness
EMMONS Brittney, UCLA 117
ERGINTAV Semih, 135
ERICKSEN Todd, 075, 192
ERICKSON Brittany, Portland State 045
EUCHNER Fabian, ETHZ 313
EVANS Eileen, USGS 147
EVANS James, Utah State 110, 116
EYMOLD William, USC 232
FAHERTY Drew, CalPoly Pomona
FAN Wenyuan, UCSD 173
FANEY Mark, 133
FANG Hongjian, 213
FARRIS Andrew, CSULB 016
FATTAHI Heresh, Caltech 076
FEAUX Karl, UNAVCO 144
FENG Tian, UCLA 230
FERDOWSI Behrooz, U Penn 038
FIALKO Yuri, SIO/UCSD 066, 129, 139,
158
FIELD Edward, USGS 298, 299, 300
FIELDING Eric, JPL 076, 157
FIGUEIREDO Paula, U Cincinnati 124, 125
FINNEMORE Mike, 084
FLESCH Lucy, Purdue 008
FLETCHER John, CICESE 073, 074
FLOURNOY James, Save Our Community
FLOYD Michael, MIT
FORD Heather, UCR
FOSDICK Julie, U Connecticut 122
FOUTZ Anna, Chaffey College
FRANCUCH Dean, Shannon & Wilson
FRANK William, 233
FRANKS Jill, USGS
FREED Andrew, Purdue 157
FROST Carol, NSF
FRYER Rosemarie, UT Austin 127
FU Yuning, Bowling Green State 012
FU Zhenyu, SCEC UseIT/USC 329
FUIS Gary, USGS 005
FULTON Patrick, TAMU 105
FUNNING Gareth, UCR 136, 151, 159,
167, 244
FURUMURA Takashi, ERI/U Tokyo 036
GABRIEL Alice-Agnes, LMU Munich 062,
341
GABRIEL Katherine, CSUN 119
GAGE Nicole, CalPoly Pomona
GAHERTY James, Columbia
GALETZKA John, Caltech 146
GALIS Martin, Comenius U in Bratislava
052
GALLOVIC Frantisek, Charles Univ
Prague 055, 288
GARCIA-REYES Jose Luis, Caltech 190
GARZA-GIRON Ricardo, UCSC 255
GEE Lind, USGS
GENTILE Christopher, UCR
WELCOME
2016 SCEC Annual Meeting page 262
GERSTENBERGER Matthew, GNS
Science Tue 14:30
GHISETTI Francesca, U Canterbury
GHOSH Abhijit, UCR 215, 239, 242, 248,
251
GILCHRIST Jacquelyn, USC 316
GILL David, SCEC/USC 327, 328, 337
GIRTY Gary, SDSU 077
GIVEN Douglas, USGS 187
GLASSCOE Margaret, JPL 138, 152, 303
GLENNIE Craig, GeoDigital Int'l 075, 192
GOEBEL Thomas, UCSC Tue 14:00
GOLD Peter, UT Austin 090, 127
GOLDBERG Dara, UCSD 256
GOLDFINGER Chris, Oregon State 113
GOLDMAN Mark, USGS 005, 081, 176,
257, 262
GOLDSBY David, U Penn 034, 038
GOLDSTEIN Ariella, SDSU
GOMEZ Luis, Chaffey College 333
GONZÀLEZ-GARCÌA Jose Javier,
CICESE 146, 155
GONZALEZ-HUIZAR Hector, UTEP 179
GOODING Margaret, LSA
GORI Marcello, Caltech 035
GORMLEY Deborah, USC/SCEC 328
GOULET Christine, SCEC/USC 002, 272,
337
GRANAT Robert, JPL 343
GRANT LUDWIG Lisa, UCI 128, 130, 343
GRAPENTHIN Ronni, New Mexico Tech
319
GRAVES Robert, USGS 022, 272, 284,
337
GREGOLO Stefano, ASISIMIC
GRIFFITH W., UT Arlington
GUALANDI Adriano, Caltech
GUAN Brady, ELAC 331
GUILLEMOT Christian, USGS 192
GUIWITS Stephan, 187
GUNS Katherine, U Arizona
GUO Hao, 120
GUTIERREZ Abel, 073
GUY Richard, Caltech 209
HAASE Jennifer, UCSD 256
HADDADI Hamid, CGS
HAFFENER Jackson, 202
HAGER Bradford, MIT 051
HAGOS Lijam, CGS
HAJAROLASVADI Setare, UIUC 041
HAJI-SOLTANI ALIREZA, U Memphis 263
HALL Craig, SDSU
HALL Kelley, UNR 259
HAMMOND William, Nevada Geodetic Lab
Tue 10:30
HANKS Tom, USGS 175
HANSON Austin, NM State 096, 099
HAO Jinglai, 235, 240
HARDEBECK Jeanne, USGS 014, 217,
298
HARDING Alistair, UCSD 109
HARDY Sandra, 179
HARRINGTON Rebecca, UCLA 197
HARRIS Cooper, USC
HARRIS Ruth, USGS 047
HARTE David, GNS Science Tue 14:30
HARTOG J, 237
HATCH Rachel, UNR 225
HATEM Alexandra, USC 017
HATFIELD Billy, UCSD 261
HAUKSSON Egill, Caltech 003, 187, 198,
218
HE Xiaohui, USTC 211
HEARN Elizabeth, Capstone Geophysics
Mon 14:00
HEATON Thomas, Caltech 276
HEDAYAT Ahmadreza, CSM 028
HEERMANCE Richard, CSUN 119, 126
HEFLIN Michael, JPL 343
HEINECKE Alexander, Tech U Munich
340, 342
HELMBERGER Donald, caltech 022, 076
HENSLEY Scott, JPL 138
HERNANDEZ Janis, CGS 081, 257
HERNÁNDEZ César, 325
HERON Chris, LADWP
HEROVIC Emina, U Kentucky 324
HERRING Thomas, MIT
HERRMAN Mike, CalPoly Pomona
HERRMANN Robert, 271
HESS Derick, 346
HICKMAN Stephen, USGS
HILL David, USGS 217
HILLEY George, UC Berkeley 111
HINOJOSA Alejandro, CICESE 073
HINTON John, 028
HIRAKAWA Evan, LLNL 069
HIRATA Naoshi, ERI/U Tokyo 273, 313
HIRTH Greg, Brown 339
HISLOP Ann, U Kentucky 102
HODGKINSON Kathleen, UNAVCO
HOGAN Phillip, Fugro
HOIRUP Don, CA DWR
HOLDEN Caroline, GNS Science 195, 275,
287
HOLLINGSWORTH James, USC 088
HOLLIS Daniel, Consultant 233
HOLMES James, UCSD 107
HOLT Bill, Stony Brook 008, 165
HOOVER Susan, 271
HOPPS Tom, Rancho Energy 006
HORNBACH Matthew, 024
HOSSAIN John, Wesleyan
HOSSEINI Mehrdad, AECOM 170
HOSSEINI Seyed, USC Tue 14:00
HOUGH Susan, USGS 171, 345
HOUSTON Heidi, U Washington 259
HU Jianping, LADWP
HU Jing, 084
HU Zhifeng, SDSU/UCSD
HUANG Emy, PCC 332
HUANG Hsin-Hua, Caltech
HUANG Hui, UCLA 160, 246
HUANG Mong-Han, JPL 157
HUANG Yihe, Stanford 193, 194
HUDA Md Monsurul, CERI 270
HUDNUT Kenneth, USGS 075, 080, 085,
145
HUDSON Kenneth, UCSB 282
HUERTA Brittany, CSUN 126
HUIBREGTSE Emma, USC 334
HUTCHISON Alexandra, UCR 242
HUYNH Tran, SCEC/USC 327, 328
HWU Wen-Mei, 344
IMPERATORI walter, ETH Zurich 288
INSERRA Nicholas, UCLA
IRWIN Tisha, UCR 159
ISHIGAKI Yuzo, JMA (Japan) 313
JACKSON David, UCLA Mon 11:00
JAGOUTZ Oliver, 339
JAHAN Naomi, Smith College 149
JANECKE Susanne, Utah State 116
JEONG Seokho, QuakeCoRE 283, 286
JHA Birendra, USC 345
JI Chen, UCSB 057, 235, 240, 258
JIANG Junle, SIO/UCSD 066
JIMENEZ Eddie, CSUF
JOHNSON Christopher, UC Berkeley 012
JOHNSON Kaj, Indiana U Tue 10:30
JOHNSON Leonard, NSF
JOHNSON Marilyn, PCC
JOHNSON Patrick, Aerospace Corp 085
JOHNSON Wade, 135
JOHNSTONE Samuel, 111
JOLIVET Romain, U Cambridge 305
JONES Lucile, Caltech
JORDAN Thomas, SCEC/USC Wed
09:00, 002, 228, 232, 236, 272, 298,
314, 315, 316, 331, 332, 333, 335, 337
JORDAN, JR. Franklin, San Bernardino
County
JUAREZ Alan, USC 234
JUAREZ Sandra, UNAM
KAGAN Yan, UCLA 301
KAISER Anna, GNS Science 287
KALKAN Erol, USGS
KAMAL Maksud, 091
KAMERLING Marc, Consultant 007
KANAMORI Hiroo, Caltech 076, 212, 243
KARIM Mir, Geocomp Corp 091
KARLSSON Keene, SDSU 073
KARLSTROM Leif, 070
KARPLUS Marianne, 201
KECK Dan, Etiwanda HS 150
KENDRICK Katherine, USGS
KENT Graham, UNR 107, 109
KHOSHMANESH Mostafa, ASU 162
KHOSHNEVIS Naeem, CERI 269, 270,
272
KILB Debi, UCSD 326
KILGORE Brian, USGS 029
KIRATZI Anastasia, 055
KIRBY Eric, Penn State 092
WELCOME
2016 SCEC Annual Meeting page 263
KIRKWOOD Robert, Riverside USD
KITAJIMA Hiroko, TAMU
KLEMPERER Simon, Stanford 113
KLINE Mark, Banning HS
KLOTSKO Shannon, UCSD
KNEPLEY Matthew, ANL 043
KNUDSEN Keith, USGS
KO Justin, Caltech
KOCH Landie, Starwood
KOHLER Monica, Caltech Tue 11:00
KONG Qingkai, UC Berkeley 249
KOTHARI Konik, UIUC 040
KOZDON Jeremy, Naval Postgrad School
045
KRANER Meredith, UNR 165, 166
KRAVITZ Katherine, UC Boulder
KREEMER Corné, UNR Tue 10:30
KRESS Victor, 237
KROLL Kayla, LLNL 317
KROPIVNITSKAYA Yelena, Western U
(Canada) 196
KRUEGER Hannah, AppState 167
KUO Szu-Ting, TAMU
KUPFERSCHMIDT Larissa, CalPoly
Pomona
KWIATEK Grzegorz, 183
KYRIAKOPOULOS Christodoulos, UCR
056, 322
LACY Amber, CalPoly Pomona 223
LAI Voon Hui, Caltech 022
LAJOIE Lia, CSM 115
LAMBECK Kurt, ANU (Australia) 083
LAMBIE Emily, GNS Science
LANDRY Walter, 148
LANGBEIN John, USGS 192
LANGRIDGE Robert, GNS Science
LAPUSTA Nadia, Caltech 035, 212
LARKIN Tam, 275
LAU Homan, UCSD/SIO 139
LAWRENCE Shawn, UNAVCO 144
LAWS Alexander, SDSU
LAY Thorne, UCSC 169, 243
LEE Robin, U Canterbury 286
LEE Yajie, ImageCat
LEEPER Robert, UCR
LEGG Mark, Legg Geophys 113
LEI Qiyun, 080
LEITH William, USGS
LEIVA Frances-Julianna, CalPoly Pomona
LEIVA Luis, 325
LEPRINCE Sebastien, Caltech 088
LEUE Nolan, 226, 252
LEVARIO Jonathan, CalPoly Pomona
LEVY Yuval, SDSU/UCSD 082
LI Bo, UCR 215, 251
LI Chenyu, Georgia Tech 214
LI Shaohua, USGS
LI Xinnan, CEA (China)
LI Yipeng, ELAC 331
LI Yong-Gang, USC 176
LI Youjuan, CEA (China)
LI Zefeng, Georgia Tech 191, 233
LI Zhanfei, 080
LIANG Cunren, 076, 140
LIEL Abbie, UC Boulder 172, 247
LIEOU Charles, UCSB 039
LIFTON Nathaniel, Purdue 099, 126
LIFTON Zach, Geosyntec
LIN Fan-Chi, U Utah 185, 199, 201
LIN Jessica, UCLA 177
LIN Ting, Marquette Wed 08:30
LIN Yen-Yu, Caltech 212
LIN Yu-Pin, USC 228
LINDSEY Eric, EOS/UC Berkeley 129
LINDVALL Scott, Lettis Consultants
LINVILLE Lisa, U Utah
LIPOVSKY Bradley, Harvard 072
LIPPOLDT Rachel, USC 253
LIU dunyu, TAMU 050
LIU Jing, CEA (China) 080, 089
LIU Lijun, 008
LIU Taojun, UC Boulder/USGS 172, 247
LIU Xin, Stanford 188
LIU Zhen, JPL/Caltech 140, 153
LIU Zhipeng, 208
LIUKIS Maria, SCEC/USC 314, 315
LLENOS Andrea, USGS 309
LO Annie, USC 328
LOCKNER David, USGS 029, 075
LOHMAN Rowena, Cornell 137
LOPEZ Hernan, SCEC UseIT 329
LOUIE John, Nevada Seismo Lab
LOVELESS Jack, Smith College 167
LOZOS Julian, CSUN 061, 323
LUCO Nicolas, USGS 172, 247
LUGINBUHL Molly, UC Davis 207
LUNA Eloy, LAFD CUPA
LUO Bin, TAMU 048
LUO Yan, Caltech 213
LUTTRELL Karen, LSU Mon 14:30, 010
LUTZ Andrew, CA DWR
LYDEEN Scott, USGS 226
LYNCH David, Spectral Sciences 085
LYZENGA Gregory, Harvey Mudd 152
MA Hongsheng, CEA (China) 020
MA Kuo-Fong, 212
MA Shuo, SDSU 069
MA Xiao, UIUC 042
MACY Kyle, CalPoly Pomona 223
MADDEN Elizabeth, UMass Amherst 015
MADUGO Christopher, PG&E 075
MAECHLING Philip, SCEC 002, 272, 315,
327, 328, 335, 337
MAGNANI Maria Beatrice, SMU 024
MAI Paul, KAUST 052
MALONEY Jillian, SDSU 086, 093
MANN Doerte, UNAVCO 144
MANN Mary, Geothermal Resource Group
MARKOWSKI Daniel, Utah State 116
MARQUEZ Eui-jo, SDSU 086
MARQUIS John, SCEC 327, 328
MARSHALL Scott, AppState 015, 151, 167
MARTIN Antony, Geovision 226, 252
MARTINEZ-BARCENA Hebert, 146
MARTÍNEZ-GARZÓN Patricia, 013, 183
MARZOCCHI Warner, INGV 314, 315
MATAGNINI Luca, 064
MATERNA Kathryn, UC Berkeley 238
MATTI Jon, USGS 108, 123
MATTIOLI Glen, UNAVCO 135, 144
MCBECK Jessica, UMass Amherst 059
MCCAFFREY Robert, Portland State 141,
142
MCDONALD Eric, CSUN 092
MCGILL Sally, CSUSB 101, 125, 149
MCGUIRE Christopher, UCLA 100
MCGUIRE Jeff, WHOI 229
MCHARGUE Tim, 111
MCKAY Nicholas, Northern Arizona U
MCNAMARA Dan, LLNL 271
MCPHILLIPS Devin, USGS 099, 226
MCRANEY John, SCEC/USC
MCVERRY Graeme, Tue 14:30
MEADE Brendan, Harvard 018, 338
MEERTENS Charles, UNAVCO Tue 10:30
MEIER Men-Andrin, Caltech 187, 210
MELBOURNE Timothy, CWU
MELGAR Diego, UC Berkeley 319
MELTZNER Aron, EOS 056, 124
MENCIN David, UNAVCO 135
MENDOZA Manuel, UCR 248, 251
MENG Haoran, USC 229
MENG Lingsen, UCLA 019, 160, 177, 208,
230
MENG Xiaofeng, Georgia Tech 214, 216
MENGES Christopher, USGS 108
MICHAEL Andrew, USGS 298, 309
MICHEL Sylvain, U Cambridge 305
MIDTTUN Nikolas, USGS 096
MILETI Dennis, UC Boulder
MILLER Kevin, CalOES
MILLER Meghan, UNAVCO
MILLINER Chris, UC Berkeley 017, 088,
152
MILNER Kevin, SCEC/USC 298, 299, 300,
331, 332, 335, 337
MINSON Sarah, USGS 075, 187, 192
MIRANDA Elena, CSUN 078
MISHIN Dmitry, 348
MOE Annelisa, CSUN
MOECHER David, 102
MOLYNEUX James, UCLA
MONTONE Paola, INGV
MOON Seulgi, UCLA 079, 117, 177
MOONEY Walter, USGS
MOORE Diane, USGS 075
MOORES Andrew, Nanometrics
MORALES Brianna, 133
MORELAN Alex, UC Davis 131
MORELAND John, 348
MOSCHETTI Morgan, USGS 271
MOSER Amy, Utah State 110
MOTAMED Ramin, UNR 281
WELCOME
2016 SCEC Annual Meeting page 264
MOYER Pamela, U New Hampshire 053
MU Dawei, UCSD/SDSC 336, 344
MUELLER Karl, UC Boulder 023
MUNOZ Juan, UT Austin 127
MURRAY Jessica, USGS 141, 154, 192
MURRAY Kyle, Cornell 137
MURRAY Lyndon, Anza-Borrego Desert
State Park
MURRAY Mark, New Mexico Tech
NADEAU David, 348
NADEAU Robert, UC Berkeley 064, 065
NAEIM Farzad, Farzad Naeim, Inc.
NAKAGAWA Shigeki, 273
NAKATA Nori, U Oklahoma 277
NANJO Kazuyoshi, Yokohama Nat U
(Japan) 313
NAZER Mohammad, QuakeCoRE 285
NEIGHBORS Corrie, UCR 231
NELSON Zachary, UCSB
NEVITT Johanna, USGS 075, 111
NG Raymond, CalPoly Pomona 206
NGU Daniel, SCEC UseIT/UCI 331
NGUYEN Rosa, CalPoly Pomona
NI Sidao, URS 211
NICHOLSON Craig, UCSB 003, 006, 007
NIE Shiying, SDSU
NIGBOR Robert, Geovision
NISSEN Ed, CSM 115
NIU Fenglin, 246
NOACK Marcus, Simula Research 349
NORMAN Michael, 348
NORMAN Rory, SCEC USEIT/GCC 332
NOVAK Jenny, CSUN 323
O'LEARY Dave, 262
O'MAHONY Luke, Trafalgar
O'REILLY Ossian, Stanford 320
OGLESBY David, UCR 056, 058
OKAYA David, USC 232
OKUMURA Koji, Hiroshima U Tue 08:30
OKUWAKI Ryo, U Tsukuba 067
OLMOS Emily, SCEC UseIT 334
OLSEN Kim, SDSU 001, 267, 272, 293,
337, 344
OLSON Brian, CGS
ONDERDONK Nate, CSULB 016
ORLANDINI Omero, UC Boulder
ORTEGA Alejandro, SIO/IGPP 146, 155,
156, 157
ORTEGA Gustavo, CALTRANS
ORTEGA Spencer, PCC 329
ORTON Alice, Geocon
OSKIN Michael, UC Davis 073, 074, 089,
131
OWEN Lewis, U Cincinnati 073, 097, 125
OWEN Susan, JPL 151, 167
OZENER Haluk, 135
PACE Alan, Petra Geosci
PADILLA Alina, Ramona Acad
PAGE Morgan, USGS Tue 14:00
PANGRCIC Hannah, 092
PARKER Grace, UCLA
PARKER Jay, JPL 138, 152, 343
PARRA Zachary, TAMU
PAUDYAL Harihar, 302
PAUK Edric, SCEC/USC 327, 328
PAULL Charles, 111
PAYNE Ryan, TAMU 049
PEARSON Carl, 344
PEARSON Jozi, SCEC/UC R 114, 331,
332, 333, 334
PEKUROVSKY Dmitry, UCSD 347
PELTZER Gilles, UCLA 079
PENG Yajun, Princeton 203
PENG Zhigang, Georgia Tech 191, 214,
216, 233
PEPPARD Daniel, SDSU 077
PEREA Hector, UCSD
PESHETTE Paul, CSUN
PETERS Robert, SDSU 087
PETERSEN Mark, USGS 307
PETERSON Matthew, U Idaho 149
PETRASHEK Stacey, CalPoly Pomona
PETTINGA Didier, 296
PEZESHK Shahram, U Memphis 263
PFAU Jennifer, 226
PHILLIPS Benjamin, NASA
PHILLIPS Wendy, FEMA
PICCIONE Gavin, 008
PIERCE Ian, Center for Neotectonic
Studies 097
PIERCE Marlon, JPL 343
PIGATI Jeffrey, 098
PLESCH Andreas, Harvard 003
PLICKA Vladislav, 055
POEPPE Dean, 037
POLAK Viktor, 350
POLET Jascha, CalPoly Pomona 206, 223,
227, 245
POLISSAR Pratigya, Columbia 025
POLLITZ Fred, USGS
PORRITT Robert, USC 253
PORTER Keith, UC Boulder 300
POWELL Robert, USGS 102
POWERS Peter, USGS 271
PREJEAN Stephanie, USGS 255
PURASINGHE Rupa, CSULA
PURASINGHE Ruwanka, LADWP
QIN Jinhui, 196
QIN Lei, USC 220
QUI Hongrui, USC 219, 220
QUIDAS Shayna, TAMU
QUINN Robert, 116
RABADE Santiago, UNAM 234
RABINOWITZ Hannah, Columbia 026
RAHMAN Zillur, 091
RAI S, 248
RAINS Christine, JPL
RAJASEKHAR Kishan, UCI 332
RAMIREZ-GUZMAN Leonardo, UNAM
234, 295
RASBURY Troy, 008
RATHJE Ellen, UT Austin 084
RAZAFINDRAKOTO Hoby, U Canterbury
284, 285, 350
RECHES Ze'ev, U Oklahoma 027
REEDY Tabor, UNR 104
REGALLA Christine, Boston U 092
REICHLIN Robin, NSF
REITMAN Nadine, UC Boulder
REMPEL Alan, U Oregon 070, 071
RENNOLET Steven, 271
RESOR Phillip, Wesleyan 015
RESTREPO Doriam, Univ EAFIT 294
RETTENBERGER Sebastian, Tech U
Munich 341
REYNERS Martin, GNS Science 021
REYNOLDS Laura, UCSB 087
REZAEIAN Sanaz, USGS
RHOADES David, GNS Science Tue
14:30, 195, 304, 308, 314, 315
RHODES Ed, UCLA 100, 101, 117
RIANO Andrea, CMU 294
RICHARDS-DINGER Keith, UCR 316,
317, 322, 347
RISTAU John, GNS Science 021
ROBLES Juan, CICESE 146
ROCKWELL Thomas, SDSU Mon 10:30,
056, 073, 082, 086, 087, 093, 124, 128,
130
ROH Becky, Caltech 276
ROLLINS Chris, Caltech 148
ROMAN Angela, 083
ROMANO Mark, Blue Tavern 334
ROSAKIS Ares, Caltech 035
ROSINSKI Anne, CGS
ROSS Zachary, Caltech 174, 218, 219, 220
ROTEN Daniel, SDSU 267, 272, 344
ROUBY Helene, ENS Paris 083
RUBIN Allan, Princeton Tue 08:00
RUBIN Ron, CGS
RUBINO Vito, Caltech 035
RUBINSTEIN Justin, USGS 181
RUDDY Brad, CalPoly Pomona 245
RUEGG Christian, 084
RUGG Scott, 086
RUHL Christine, UC Berkeley 319
RUNDLE John, UC Davis 207, 303, 318,
343
RYAN Kenny, UCR
SABALA Luke, US NPS 102
SABBETH Leah, Caltech
SABER Omid, TAMU 031, 032
SAHAKIAN Valerie, USGS 175
SALAS Valeria, 325
SALAZAR MONROY Edilson, UNAM 295
SALISBURY James, ASU 130
SAMMIS Charles, USC 253
SANCHEZ Robert, 209
SANCHEZ-SESMA Francisco, 295
SANDOW Sharon, SCEC
SANDWELL David, UCSD 010, 134, 141,
155, 156
SANSKRITI FNU, SCEC 332, 335
WELCOME
2016 SCEC Annual Meeting page 265
SANTIBÁÑEZ Isabel, 325
SARE Robert, Stanford 111
SAUNDERS Jessie, UCSD 256
SAVAGE Heather, Columbia 025, 026
SAVARD Genevieve, Tue 08:00
SAVRAN William, SDSU 272, 293, 344
SAWYER Thomas, 094
SAY Michael, UCLA
SCHARER Katherine, USGS 085, 096,
099, 118, 121
SCHEIRER Daniel, USGS 005
SCHMIDT David, U Washington 259
SCHOENBALL Martin, Stanford 183
SCHORLEMMER Danijel, GFZ Potsdam
312, 313, 315
SCHREIER Louis, 249
SCHULTE-PELKUM Vera, UC Boulder
023
SCHULTHIES Sarah, 110
SCHULTZ Kasey, UC Davis 318
SCHUSTER Gerard, U Utah 201
SCHWARTZ David, USGS
SCROGGINS Kevin, PCC 329
SEGALL Paul, Stanford 044, 161
SEITZ Gordon, CGS
SELLNOW Deanna, U Central Florida 324
SELLNOW Timothy, U Central Florida 324
SENOBARI Nader, UCR 244
SERPETSIDAKI Anna, 055
SESSIONS Alex, 311
SEVERINO RIVAS Vianca, SCEC UseIT/U
Puerto Rico 333
SHANKER Prof Daya, Indian Ins Tech
Roorkee 302
SHAO Yanxiu, CEA (China) 080, 089
SHAO Zhigang, CEA (China) 020
SHARE Pieter-Ewald, USC 120, 185, 201,
219, 220
SHARP Warren, Berkeley Geochron 127
SHAUGHNESSY Liam, IRIS 231
SHAW Bruce, Columbia 193, 298, 347
SHAW John, Harvard 003
SHEARER Peter, UCSD 173, 200, 205,
222, 225
SHELLY David, USGS Tue 08:00
SHEN Xueyao, USC
SHEN Zheng-Kang, UCLA 140, 141, 307
SHENG Yixiao, Stanford 274
SHI Allen, USC 332
SHI Jian, Caltech 266
SHI Zheqiang, SDSU 272
SHINEVAR William, MIT/WHOI 339
SHIRZAEI Manoochehr, ASU 162, 163
SHUMAKER Lauren, 111
SIBSON Richard, U Otago Sun 18:00
SICKLER Robert, USGS 081, 257, 262
SIDDIQUA Sumi, 091
SILVA Fabio, SCEC/USC 002, 272, 292
SIMAN-TOV SHALEV, UCSC 054
SIMILA Gerry, CSUN
SIMMS Alexander, UCSB 083, 087, 106
SIMONS Mark, Caltech 076
SINGLETON Drake, SDSU
SKARBEK Robert, Columbia 071
SKARLATOUDIS Andreas, AECOM 170,
292
SKOUMAL Rob, Stanford 189
SLADEN Anthony, Caltech 070
SLEEP Norman, Stanford 279
SMITH Deborah, USGS
SMITH Ken, Nevada Seismo Lab 200, 225
SMITH Rubin, 008
SMITH Trevor, UCSB 258
SMITH-KONTER Bridget, U Hawaii 010,
156, 179
SOKOS Efthimios, 055
SOMERVILLE Paul, AECOM 170, 292
SONG Teh-Ru, Caltech 212
SONG Xin, USC 236
SORLIEN Christopher, UCSB 003, 006,
007
SPAGNUOLO Elena, 026
SPARKS David, TAMU 049
SPELZ Ronald, UNAM 073
SPICA Zack, Stanford 186
SPINLER Joshua, U Arkansas 149
SPRINGER Kathleen, USGS 098
STARK Keith, Strata Info Tech 145
STEIDL Jamison, UCSB
STEPHENSON Ollie, Caltech
STERN Aviel, UMass Amherst 058
STEWART Craig, CSUN 078
STEWART Jonathan, UCLA
STIRLING Mark, U Otago 290
STOCK Greg, 054
STOCK Joann, Caltech 085
STUPAZZINI Marco, Munich Re
SU Feng, USBR
SUMY Danielle, IRIS 231
SUN Xiaolong, CEA (China)
SUTKOWSKI Chloe, CalPoly Pomona 227
SUTTON Jessica, ASU
SVARC Jerry, USGS 154
SWANSON Brian, CGS
SWIATLOWSKI Jerlyn, UCR 136
SWINDLE Carl, UCSB 132
SYLVAIN Michael, 076
TABORDA Ricardo, U Memphis 266, 269,
270, 272, 294
TAHIR Zain, SDSU
TAIRA Taka'aki, UC Berkeley 064, 238
TAKEDATSU Rumi, SDSU 001
TAL Yuval, MIT 051
TANIMOTO Toshiro, UCSB 224
TARNOWSKI Jennifer, UCR 058, 239
TARONI Matteo, INGV 314
TENG Ta-liang, USC
TERAN Orlando, CICESE 074
THATCHER Wayne, USGS 011
THIO Hong Kie, AECOM 170
THOM Christopher, U Penn 034
THOMAS Amanda, U Oregon Tue 08:00,
070
THOMAS Valerie, USGS 198
THOMPSON Eric, SDSU 271
THOMPSON Thomas, Harvard 018, 338
THURBER Clifford, UW Madison 120, 213
TIAMPO Kristy, UC Boulder 164, 196
TOBIN Josh, UCSD 336, 340
TOKE Nathan, Utah State
TONG Xiaopeng, U Washington 010, 156
TRATT David, Aerospace Corp 085
TREIMAN Jerry, CGS
TRUGMAN Daniel, UCSD 200, 222, 225,
326
TSAI Victor, Caltech 171, 212, 345
TSANG Rebecca, SDSU 124
TSURUOKA Hiroshi, ERI/U Tokyo 313
TULLIS Terry, Brown 030
TURCOTTE Donald, UC Davis 207
TURNER Ryan, UNAVCO 144
TWARDZIK Cedric, UCSB 240
TYLER mitch, CA DWR
TYMOFYEYEVA Ekaterina, UCSD 139
TYSON Jeffrey, LADWP
UCARKUS Gülsen, Istanbul Tech U
UCHIDE Takahiko, GSJ/NAIST 312
UPHOFF Carsten, Tech U Munich 341
VAHI Karan, USC 337
VALLIN Laura, 073
VALOVCIN Anne, UCSB 224
VAN DER ELST Nicholas, USGS Tue
14:00
VAN DER HILST Rob, MIT 204, 246
VAN DER WOERD Jerome, 089
VAN DISSEN Russ, GNS Science 100
VAN HOUTTE Chris, Harvard 275, 290
VANDERWALL Patrick, SCEC UseIT 332
VARGAS Bernadette, Etiwanda HS
VERNON Frank, UCSD 174, 185, 201,
219, 220
VIDALE John, U Washington 237
VIESCA Robert, Tufts 068
VILLAVERDE Víctor, CICESE 073
VINCI Margaret, Caltech
WADE adam, ASU
WALKER Robert, USC 171, 345
WALLACE Laura, UT Austin 021
WALLS Christian, UNAVCO 144, 146
WANG Jiong, UCSB 224
WANG Jun, Indiana U 343
WANG Kang, UCSD 158
WANG Lifeng, GFZ Potsdam 305
WANG Nan, SDSU/SIO
WANG Peng, 089, 344
WANG Pengtao, 089
WANG Wang, 089
WANG Wei, UCSD 205
WARD Steven, UCSC
WEBB Heather, SDSU 077
WEI Shengji, Caltech 022
WEIDMAN Luke, SDSU 093
WELCOME
2016 SCEC Annual Meeting page 266
WEINGARTEN Matthew, USGS 202
WEISER Debbie, USGS Tue 14:00
WELCH Drew, SCEC 334
WELLS Jack, 116
WERNER Maximilian, Bristol 298, 314, 315
WERSAN Louis, Indiana 122
WESNOUSKY Steven, UNR 094, 095,
097, 104
WEST Josh, USC 311
WETZLER Nadav, UCSC 169
WHITCOMB James, NSF
WHITE Elizabeth, SDSU
WHITE Isabel, USGS 247
WHITE Joshua, 317
WHITE Malcolm, USC 174
WHITTAKER Andrew, MCEER/U Buffalo
WIGGINTON Sarah, Utah State
WILLIAMS Alana, ASU 128
WILLIAMS Charles, GNS Science 021,
043, 195
WILLIAMS Colin, USGS 011
WILLIAMS Patrick, Williams Assoc
WILLIS Michael, Cornell 133
WILLS Chris, CGS
WILSON Andrew, SDSU
WILSON John, UC Davis 303, 318
WITHERS Kyle, SDSU 178, 344
WITKOSKY Ryan, Caltech 085
WOLFE Cecily, USGS
WOLIN Emily, USGS
WOLLHERR Stephanie, LMU Munich 062
WOODDELL Katie, PG&E 064
WORKING GROUP the CSEP, 315
WU Qimin, Virginia Tech 241
WU Yanqiang, 143
WU Zhongliang, CEA (China)
WYATT Frank, UCSD 261
XIE Yuqing, UCLA 019
XU Jing, 020
XU Xiaohua, UCSD/SIO/IGPP 141, 168
XUE Lian, UCSC 105
YAGI Yuji, U of Tsukuba (Japan) 067
YAO Dongdong, Georgia Tech 214, 216,
233
YAO Huajian, UCSD 246
YAO Wenqian, 089
YAO Zhenxing, 235
YE Lingling, Caltech 243
YECK William, 271
YEH Te-Yang, SDSU
YILMAZ Onur, 135
YIN Gregory, UCLA
YODER Mark, UC Davis 303
YONG Alan, USGS 226
YONG Zhang, CEA (China)
YOON Clara, Stanford 194
YOSHIMITSU Nana, Stanford 036, 193
YOUNG Elaine, UC Davis 118
YOUNG Holly, 111
YOUNG Jerry, Cielo Vista Charter
YOUNG Karen, USC/SCEC
YOUNT Charles, 340
YU Chunquan, Caltech 204
YU Ellen, Caltech 198
YU John, USC 315, 327, 328, 332, 335
YUE Han, Caltech 076
YULE Doug, CSUN 119, 126
YUN Sang-Ho, JPL
ZAHRADNIK Jiri, Charles Univ Prague 055
ZALIAPIN Ilya, UNR 306, 310
ZAMORA Jenepher, PCC 333
ZANZERKIA Eva, NSF
ZAREIAN Farzin, UCI
ZECHAR Zechar, ETHZ 314, 315
ZEIDAN Tina
ZENG Xiangfang, U Wisconsin 213
ZENG Yuehua, USGS 141, 307
ZHAN Zhongwen, Caltech
ZHANG Dandan, CalPoly Pomona/Caltech
ZHANG Edward, 005
ZHANG Haijiang, UW Madison 120, 213
ZHANG Jinyu, 089
ZHANG Langping, 020
ZHANG Peizhen, CEA (China) 080
ZHANG XIAODONG, CEA (China)
ZHANG Yongxian, CEA (China)
ZHAO Yonghong, 019
ZHONG Peng, UCI
ZHOU Hui, CEA (China) 336
ZHOU Mingyan, SCEC 331
ZHU Linjun, 076
ZIELKE Olaf, KAUST 052
ZINKE Robert, USC 088, 100
ZU Ximeng, U Oklahoma 027
ZUMBERGE Mark, UCSD 261
ZURBUCHEN Julie, UCSB 106
2016 SCEC Annual Meeting page 267
Center Management Administration Education & Outreach Information Technology
Center Director Tom Jordan
Associate Director John McRaney
Associate Director Mark Benthien
Associate Director Phil Maechling
Co-Director Greg Beroza
Special Projects/Events Tran Huynh
Strategic Partnerships Sharon Sandow
Research Programmer Scott Callaghan David Gill Masha Liukis Kevin Milner Edric Pauk Fabio Silva Systems Programmer John Yu
Exec Dir Spec Proj Christine Goulet
Contracts & Grants Karen Young
Digital Products John Marquis
Admin Coordinator Deborah Gormley
Communications Jason Ballmann
Core Institutions and Representatives USC, lead Tom Jordan
Harvard Jim Rice
Texas A&M Patrick Fulton
UC Santa Barbara Ralph Archuleta
USGS Menlo Park Ruth Harris and Steve Hickman
Caltech Nadia Lapusta, VC
MIT Tom Herring
UC Los Angeles Peter Bird
UC Santa Cruz Emily Brodsky
USGS Pasadena Rob Graves
CGS Chris Wills
SDSU Tom Rockwell
UC Riverside David Oglesby
UNR Glenn Biasi
At-Large Member Roland Bürgmann
Columbia Bruce Shaw
Stanford Paul Segall
UC San Diego Yuri Fialko
USGS Golden Jill McCarthy
At-Large Member Michele Cooke
Domestic Participating Institutions and Representatives Appalachian State Scott Marshall
CSU San Bernardino Sally McGill
Oregon State Andrew Meigs
UC Berkeley Roland Bürgmann
U Oregon Ray Weldon
Arizona State J Ramon Arrowsmith
Carnegie Mellon Jacobo Bielak
Penn State Eric Kirby
UC Davis Michael Oskin
U Texas El Paso Bridget Smith-Konter
Boston University Christine Regalla
Colorado Sch. Mines Edwin Nissen
Portland State Brittany Erickson
UC Irvine Lisa Grant Ludwig
U Texas Austin Whitney Behr
Brown Terry Tullis
Cornell Rowena Lohman
Purdue Andrew Freed
U Cincinnati Lewis Owen
U Wisconsin Madison Clifford Thurber
CalPoly Pomona Jascha Polet
Georgia Tech Zhigang Peng
Smith John Loveless
U Illinois Karin Dahmen
URS Corporation Paul Somerville
CSU Fullerton Dave Bowman
Indiana Kaj Johnson
SMU M. Beatrice Magnani
U Kentucky Sean Bemis
Utah State Susanne Janecke
CSU Long Beach Nate Onderdonk
JPL Andrea Donnellan
SUNY at Stony Brook William Holt
U Massachusetts Michele Cooke
Utah Valley Nathan Toke
CSU Northridge Doug Yule
LLNL Arben Pitarka
Tufts Robert Viesca
U Michigan Ann Arbor Eric Hetland
Utah Valley Nathan Toke
CSU Sacramento Steve Skinner
Marquette U Ting Lin
U Alaska Fairbanks Carl Tape
U New Hampshire Margaret Boettcher
WHOI Jeff McGuire
International Participating InstitutionsAcademia Sinica (Taiwan) ERI Tokyo (Japan) Nat’l Central U (Taiwan) U Canterbury (New Zealand)
Brendon Bradley CICESE (Mexico) ETH Zürich (Switzerland) Nat’l Taiwan U (Taiwan) U Otago (New Zealand)
Mark Stirling DPRI Kyoto (Japan) GNS (New Zealand) U Bristol (United Kingdom)
Max Werner Western Univ (Canada)
PLAZA BALLROOM
HARVEY’SBAR
RESTAURANTTAPESTRYROOM
LOBBY
BOARDROOM
Main Entrance
POOL
HORIZON BALLROOM(2nd floor)
POOL BAR
Nor
th C
alle
Enc
ilia
East Tahquitz Canyon Way
N
HILTON PALM SPRING RESORT400 East Tahquitz Canyon WayPalm Springs, CA 92262(760) 320-6868
PALM CANYON A
PALM CANYON B
WHITEWATER
(2nd floor)
OASIS II 2nd floorOASIS III 3rd floor
SATURDAY, September 1009:00-17:00 Workshop: SoSAFE (Palm Canyon)09:00-17:00 Workshop: Ventura SFSA (Plaza AB)09:00-17:00 Workshop: Fault Mechanics (Plaza CD)
SUNDAY, September 1107:00-17:00 Registration and Check-In (Lobby)08:00-12:00 Workshop: CISM (Horizon)13:00-17:00 Workshop: GMSV (Horizon)13:00-17:00 Workshop: CEO (Palm Canyon)15:00-20:00 Poster Set-Up (Plaza)17:00-18:00 Annual Meeting Welcome Social (Lobby, Harvey’s, Plaza)18:00-19:00 Distinguished Speaker Presentation (Horizon)19:00-20:30 Welcome Dinner (Poolside)19:00-21:00 SCEC Advisory Council Dinner Meeting (Palm Canyon)21:00-22:30 Poster Session (Plaza)
MONDAY, September 1207:00-08:00 Registration and Check-In (Lobby)10:00-15:00 Registration and Check-In (Lobby)07:00-08:00 Breakfast (Poolside)08:00-10:00 Session: The State of SCEC (Horizon)10:30-12:30 Session: Modeling Fault Systems-Supercycles (Horizon)12:30-14:00 Lunch (Restaurant, Tapestry, Poolside)14:00-16:00 Session: Modeling Fault Systems-Community Models (Horizon)
MONDAY, September 12 (continued)16:00-17:30 Poster Session (Plaza)19:00-21:00 SCEC Honors Banquet (Hard Rock Hotel Ballroom)21:00-22:30 Poster Session (Plaza)
TUESDAY, September 1307:00-08:00 Breakfast (Poolside)08:00-10:00 Session: Understanding Eartquake Processes (Horizon)10:30-12:30 Session: New Observations (Horizon)12:30-14:00 Lunch (Restaurant, Tapestry, Poolside)14:00-16:00 Session: Characterizing Seismic Hazard-Operational Earthquake Forecasting (Horizon)16:00-17:30 Poster Session (Plaza)19:00-21:00 Dinner (Poolside)21:00-22:30 Poster Session (Plaza)22:30-23:00 Poster Removal (Plaza)
WEDNESDAY, September 1407:00-08:00 Breakfast (Poolside)08:00-08:30 SCEC Advisory Council Report (Horizon)08:30-10:30 Session: Reducing Seismic Risk (Horizon)10:30-12:00 Session: The Future of SCEC (Horizon)12:00 Adjourn 2016 SCEC Annual Meeting
12:30-14:30 SCEC PC Lunch Meeting (Palm Canyon)SCEC Board Lunch Meeting (Tapestry)